About the Author

Michael Benton is Professor ofVertebrate Palaeontology and waspreviously also Head of the Departmentof Earth Sciences at the University ofBristol. His many books includestandard reference works and textbooksas well as popular books aboutdinosaurs and the history of life.

Other titles by Michael J. Bentonpublished by

Thames & Hudson include:

The Seventy Great Mysteries of theNatural World

Other titles of interest published byThames & Hudson include:

The Complete Ice Age: How ClimateChange Shaped the World

The Reign of the Dinosaurs

The Story of Fossils: In Search ofVanished Worlds

Our Fragile World: The Beauty of aPlanet Under Pressure

Our Living Earth

The Great Naturalists

Elephant!

The Illustrated Natural History ofSelborne

See our websiteswww.thamesandhudson.com

www.thamesandhudsonusa.com

To Mary, Philippa and Donaldfor their forbearance and inspiration,as ever

ACKNOWLEDGMENTS

I am hugely grateful to Colin Ridler of Thames& Hudson for first suggesting that I shouldwrite this book, and then for discreetly guidingme through various false starts. Thanks to him,and to his team, for their thorough andextensive work on the book as it went throughthe publication process. The entire text hasbeen read by Paul Pearson and RichardTwitchett at the University of Bristol, JamesLovelock at the University of Oxford, and PaulWignall at the University of Leeds. I value thetime they spent reading the draft manuscript,their wise suggestions about style and content,but especially for picking up some glaringmisunderstandings and errors of fact. Thanksalso to archivists and librarians at theGeological Society of London, the BritishLibrary and the Natural History Museum, for

finding ancient letters and notebooks byMurchison, Owen and the others. Finally, theappearance of the book has been enhancedimmeasurably by the artwork, and I thank JohnSibbick from Bath for his exquisitecraftsmanship informed by a real knowledge ofthe ancient beasts he draws. Finally, I dedicatethis edition to Vitaly Georgievich Ochev, ourdelightful companion in the field in Russia(chapter 10), who died in 2004.

CONTENTS

PROLOGUE

THE GEOLOGICAL TIME-SCALE

Chapter 1ANTEDILUVIAN SAURIA

Chapter 2MURCHISON NAMES THE PERMIAN

Chapter 3THE DEATH OF CATASTROPHISM

Chapter 4THE CONCEPT THAT DARED NOT SPEAK ITSNAME

Chapter 5IMPACT!

Chapter 6DIVERSITY, EXTINCTION AND MASSEXTINCTION

Chapter 7HOMING IN ON THE EVENT

Chapter 8LIFE’S BIGGEST CHALLENGE

Chapter 9A TALE OF TWO CONTINENTS

Chapter 10ON THE RIVER SAKMARA

Chapter 11WHAT CAUSED THE BIGGEST CATASTROPHEOF ALL TIME?

Chapter 12THE SIXTH MASS EXTINCTION?

GLOSSARY

NOTES

BIBLIOGRAPHY

ILLUSTRATION CREDITS

INDEX

COPYRIGHT

PROLOGUE

When I was a student of palaeontologyin the late 1970s, I remember beingpuzzled by some extraordinarydifferences of view. These were notminor disagreements – that sort of thingis common enough, and such disputes areusually resolved eventually by thediscovery of some new evidence thatsettles the matter one way or the other.This dispute was over the biggest massextinction of all time: somepalaeontologists accepted that, around

250 million years ago, all of life onEarth came very close to completeannihilation, while another group arguedthat nothing at all had really happened.How could there be any debate aboutsuch a fundamental question? The fossilrecord of the history of life may bepatchy and full of holes, but surely thereshould be no doubt about such an event?Either it had happened or it hadn’t.

As I delved deeper into the question,I was convinced that there had indeedbeen a mass extinction of hugemagnitude about 250 million years ago,at the end of the Permian period. Butwhy had some palaeontologists deniedit? The extinction deniers were notcreationists, flat-earthers or members of

some other fringe group. They knewtheir fossils, and yet their reading of therecord seemed to tell them that the endof the Permian passed with only themerest blip, the smallest disturbance,and really it was nothing to beconcerned about.

The story has moved on now. Wherethere was lack of clarity in 1970, oreven 1990, there are now two stronglyargued catastrophic models for the end-Permian mass extinction. One ties theevent to massive volcanic eruptions,producing thousands of cubic kilometresof lava and poisoning the atmosphere.The other links it to impact – dramaticnew evidence published in 2001suggests that a huge meteorite hit the

Earth and caused global destruction.Such an extraterrestrial propositionwould have been treated as wild scare-mongering by my geology professors 25years ago, and yet today it is debatedseriously.

The end-Permian mass extinction isvery much a live issue. But it is notmerely a question of interest topalaeontologists, the scientists whostudy fossils and the history of life, andto juvenile dinosaur fanatics. This is notan arcane debate that might yield only afew amusing historical or philosophicalinsights. There is immediate relevancetoday. It has often been claimed that weare living through a mass extinction atthe moment, where the main agent of

destruction is human activity. Whateverthe causes and outcomes of the currentecological crisis, the only analogues wehave for realistic comparison areextinction events in the fossil record.They can give indications of what might,or might not, happen in the future.

Death of the dinosaurs

In the late 1970s, much more waswritten about the death of the dinosaurs,the mass extinction that happened 65million years ago, at the end of theCretaceous period. But even for thisheavily studied event, there was hugedisagreement. The common view was

that the dinosaurs had dwindled toextinction gradually over perhaps 5million years. Certainly, from time totime, rather wild proposals had been putforward – that there had been a hugesolar flare, a massive meteorite impactor some other extraterrestrialcatastrophe – but these were allapparently nonsense.

Geologists in 1970 knew that theEarth was subject to huge forces thatmoved the continents at a slow pace, butmassive impacts? Not likely. Then, in1980, came one of the most astonishingpublications of my lifetime. A group ofphysicists and geologists in Californiaannounced evidence that the Earth hadindeed been hit by a huge meteorite 65

million years ago, and that theconsequences of this impact had wipedout much of life, including the dinosaurs.Their idea was met by instant ridiculeand derision by most geologists andpalaeontologists. And yet it is broadlyaccepted now.

This marks one of the biggest shiftsin scientific opinion of recent decades.From being regarded as pariahs, thecatastrophists, geologists who point tolarger-than-normal crises in thegeological past, have won the argument,in terms of extinctions of life in the pastat least. In retrospect now, it isextraordinary to see how mainstreamgeologists denied the reality ofcatastrophes for so long. Their stance

dates back to the 1830s, and earlydebates in geology which were won byt h e uniformitarians, scientists whoargued that everything in the past couldbe explained by reference to modernslow-moving processes. As a result, itbecame pretty well impossible todiscuss mass extinctions as we nowunderstand them, without being brandeda madman.

Now that it is considered acceptableto talk about sudden extinction events inthe past, the end-Cretaceous massextinction has been more closely studiedthan any other. Dozens of scientificpublications on the subject appear eachyear, and these are accompanied byhundreds of treatments in children’s

books, news items and web sites. This ispartly because of the appeal of thedinosaurs: everyone wants to know whythey died out. But also, the end-Cretaceous event was the last of severalmass extinctions, and so it is easier tostudy. It’s a general principle of geologyand palaeontology that the quality ofinformation declines the further back intime one goes. An event that happened65 million years ago is easier to dateaccurately – and associated fossils aremore abundant – than an event thathappened 251 million years ago.

Near-annihilation

The end-Permian mass extinction may beless well known than the end-Cretaceous, but it was by far the biggestmass extinction of all time. Perhaps asfew as 10% of species survived the endof the Permian, whereas 50% survivedthe end of the Cretaceous. Fifty percentextinction was clearly significantenough, and it was associated withdevastating environmental upheaval. Thecatalogue of crises 65 million years agomakes terrifying reading: the impact of avast meteorite in the Caribbean,vulcanism in India, major environmentaldeterioration, dust clouds, blacking-outof the sun, freezing cold, acid rain, anddinosaurs going extinct on all sides. Butthere is an enormous difference between

50% survival through the end-Cretaceous crisis and only 10% survivalthrough the end-Permian.

The key difference between 50%survival and 10% survival is thediversity of founders available for thereflowering of life after the catastrophe.Fifty per cent of species should offer areasonable cross-section of the range oforganisms that existed before the event,and the chance that the ecology ofhabitats on land and in the sea couldeventually recover to something like abalanced system. But survival of only10% of species means that many majorgroups of plants, animals and microbeshave probably gone forever.

Life can be thought of best as a great

tree, originating with a single foundingspecies many billions of years ago andcontinually branching and expandingupwards through time as new speciesarise. Here and there twigs of the treedie off as species become extinct, but theoverall shape of the tree expands everupward. During a mass extinction, vastswathes of the tree are cut short, as ifattacked by crazed, axe-wieldingmadmen. Whole branches and twigs arebrutally removed. The ragged remnantsof the tree, though, will reshapethemselves and grow back to fullluxuriance. At the end of the Permian,however, the slashing of the tree of lifewas vicious and sustained. Entireregions of its diverse branches were cut

and hacked off. After the crisis, only 10in 100 of the branches remained, apathetic remnant. There was certainly noguarantee that these would be sufficientto survive in the long term. After such asevere attack, the great tree of life, withover 3000 million years of historybehind it at the time, might havewithered away and died completely.

The fact that you are reading thisbook is evidence that this did nothappen, of course. The 10% of survivingspecies after the end-Permian crisisclearly were sufficient to allow thediversity of life to recover. But it was anastonishingly slow process, and therecovery phase lasted much longer thanany other known post-extinction

recovery phase.But what kind of environmental

crisis would be sufficient to wreak suchhavoc? Remember that palaeontologistshere are not looking for some processthat might have killed a few large ormarginal animals, the Permianequivalents of dodos, tigers or pandas.They are looking for a process, orcombination of processes, that wouldkill 90% of species of plants andanimals in the sea and on land, as wellas, presumably, 90% of microscopicorganisms in all habitats.

Impact

The New York Times for 23 February2001 reported ‘Meteor crash led toextinctions in era before dinosaurs’. TheTimes of London for the same day said,‘Asteroid collision left the world almostlifeless’. The Daily Mail went further inannouncing ‘The great dying. This weekone of the Earth’s greatest mysteries wassolved: What happened 250 millionyears ago to wipe out life on the globeand usher in the dinosaurs.’ A team ofscientists from the University ofWashington, NASA and otherinstitutions, claimed they had foundunequivocal proof that the end-Permianmass extinction had been caused by theimpact of a meteorite. But is the problemreally solved?

Certainly, the news reports wereconfident enough. The researchersestimated that an asteroid, or giantmeteorite, some 6 to 12 kilometresacross, had hit the Earth, and that thishad led to a series of environmentalshocks that in turn caused the extinctionsin the sea and on land. Luann Becker, theleading scientist of the group, wasreported as saying, ‘To knock out 90percent of organisms, you’ve got toattack them on more than one front’.

The significance of this event washighlighted on many web sites. One ofthem1 gave some graphic commentsabout what followed the impact:

Earthquakes and volcanoes would have rattled

the planet.… Lava poured out in volume –enough to cover the planet 10 feet (3 metres)deep. The oceans dropped 820 feet (250metres).

Worse, the combined effects of the objectvaporizing on impact, along with all thevulcanism, poisoned what was left of the seasand choked the air with ash and deadly gases.Sunlight may have disappeared for months.

The researchers say that the volcanicactivity was likely going on before the impact,but was then fuelled into a frenzy. The one-twopunch, it seems, may be what’s needed toprecipitate the worst extinctions. Poreda [oneof the contributing scientists] called the wholescenario a ‘blast from a double-barrelledshotgun’.

The picture, then, seemed clear, inFebruary 2001. The biggest mass

extinction all time, the end-Permian, hadbeen caused by a sequence of events ofunimaginable ghastliness, triggered byan extraterrestrial impact. And thatwasn’t all.

Following from this scenario, aNASA scientist was quoted as saying,‘This suggests that the evolution of lifeon Earth is strongly coupled with ourcosmic environment’. He thought that‘the Earth’s biosphere is regularlydisrupted, every 100 million years or so,by giant impacts that would renderhuman life impossible.’ So when willthe next one happen?

It’s not just wild-eyed eccentricswho scan the skies, fearing ourimpending annihilation as a result of the

impact of another such asteroid.Politicians and scientific advisers inmany countries are planning for thiseventuality. Money is being spent to fundastronomers who search for asteroidsthat might approach the Earth.Politicians warn that we must developtechnologies that would allow us to blastany asteroid that looked dangerous.Fortunately, the man from NASA assuresus that ‘there’s no need to panic, and noneed right now to spend money to defendEarth against any potential species-ending impact.’ He goes on to explainthat, if major impacts happen about onceevery 100 million years, then we mightbe able to hold on for a few moredecades, or even centuries, without too

much problem.This theory of impact has been hotly

rejected by many, or most, researchersthrough 2001 and 2002. They fight hardfor the alternative model, that massiveeruptions of lava, sustained over half-a-million years or more, causedcatastrophic environmental deterioration– poison gas, heating, freezing, strippingof soils and plants from the landscape,eruption of gases from their frozenlocations deep in the oceans and massde-oxygenation. They debate thesynergistic effects of such a break-downin earth systems, when normal feedbackprocesses are overwhelmed. We haveoceanic burps and runaway greenhouses.The volcanic model has been developed

step-by-step through the 1990s, and ithas progressively picked up adherentsand enthusiasts. So who is right – wasthe biggest mass extinction of all timetriggered by volcanic eruptions or bymeteorite impact? This book sets out toprovide the answer.

The ripples from ancient massextinctions spread a long way. The end-Permian catastrophe is clearly no longera minor sideline in the history of theEarth, as I was taught by my geologyprofessors. Perhaps nothing so familiaras the dinosaurs was killed off, but theevent was profoundly more important.Scientific findings about the events of251 million years ago can inform ourunderstanding of modern extinction

threats in a direct way.

When life nearly died

My personal involvement in exploringthe end-Permian mass extinction took meto Russia several times in the 1990s.Shortly after glasnost and perestroikawe had the opportunity to mountexpeditions to the Ural Mountains, theborderland between Europe and Asia,where some of the best Permian rocksare to be seen. We were looking at thesuccessions of ancient environmentsthere, and the evolution of amphibiansand reptiles on land. In a way, too, wewere recreating history. The Permian

period was named in 1841 after the cityof Perm on the western edge of the UralMountains, and it was named by a rovingScotsman. Sir Roderick ImpeyMurchison, one of the most distinguishedgeologists of his day, was engaged in arapid and gruelling tour of Russia duringwhich he made the first studies of theclassic Permian, and his early role is acritical part of our story.

In this book, we shall exploreseveral themes. We meet Sir RichardOwen, and some other Victorianworthies, who were the first to explorethe fossil reptiles of the past, includingthe dinosaurs. These were early days forpalaeontology, and, with newdiscoveries coming in from all corners

of the world, they had to grapple withthe astonishing concept of hugeantediluvian reptiles, much larger thanany modern lizard or crocodile. At thesame time, geologists were disentanglingthe sequence of the rocks. They werelaying the foundations of ourunderstanding of geological time anddating. This story includes a journeythrough the gentlemen’s clubs of Londonin the late 1830s, Murchison’s statelyprogress through Russia in 1840 and1841, and his evidence for naming thePermian period.

Then, with the evidence in place, weshall explore in Chapter 3 why many, ifnot most, geologists and palaeontologistsdenied the reality of a substantial crisis

at the end of the Permian. In parallelwith this denial was the problem ofunderstanding the demise of thedinosaurs, and other extinctions. Someextraordinary ideas were put forwardright up to the 1970s.

The phase of denial of the end-Permian mass extinction has ended onlyrelatively recently. Chapters 4 and 5explore the two major developments thatforced this shift of opinion: therediscovery of catastrophism –‘neocatastrophism’ – and the acceptancethat the Earth has undergone unexpectedand huge upheavals in the past. Thechange of opinion was crystallized inthat single remarkable publication in1980 that offered a robust presentation

of evidence that the dinosaurs had diedout as a result of a massive meteoriteimpact. Through this time of adjustment,from about 1950 to 1990, progress wasslow: many geologists found itimmensely hard to accept such wild andpreposterous kinds of processes in Earthhistory.

Then in Chapter 6 we explore thehistory of life, and how it has diversifiedfrom a single species to many millionstoday. Several mass extinctions, andmany smaller extinction events, havepunctuated the evolution of life fromtime to time. What has been the role ofextinction, and are there any general orpredictable patterns?

Next, we look at the end-Permian

mass extinction in detail. To do this, wefollow geologists on their recentexplorations of the evidence in China,Russia, Pakistan, Italy, Greenland andother parts of the world, as they strive topin down the exact timing of the event(Chapter 7). Their work has alsorevealed some startling new evidenceabout the precise sequence ofenvironmental changes, before, duringand after the crisis. The latest researchon the mass extinctions in the sea(Chapter 8) reveals a great deal of detailabout the magnitude of the event, whichgroups died out and which survived,whether the extinctions wereecologically selective or not, andwhether they happened equally in the

tropics and at the poles.Dramatic new discoveries also tell

us much about what was happening onland at the end of the Permian. Intenselydetailed studies of ancient soils, riversands and lake muds have revealed thechanges in plant life and in theamphibians and reptiles that fed on theplants, and on each other (Chapter 9).These studies also show what the post-apocalyptic scene was like. After themass dyings, the Earth was a cold,gloomy place, and the few survivingplants and animals experienced astrange, empty world. Our work inRussia (Chapter 10) contributes newevidence concerning the extinctions oflife on land, and in particular the

extinctions of large reptiles.Drawing all the threads together –

the evidence of ancient environmentsread from the rocks, the evidence formajor climatic changes read fromchemicals within those rocks and thedetailed record of appearances anddisappearances of microbes, plants andanimals on land and in the sea – leads toa cohesive picture of just what happenedat the end of the Permian (Chapter 11).The debate about meteorite impact vs.volcanic eruption continues apace, and Ihighlight the strengths and weaknesses ofeach theory. In the end, there does seemto be a clear winner, one whichprovides a convincing model of thesequence of killings. This is a model that

could not have been known, andcertainly would not have been believed,30 years ago. Indeed, it’s a model thatcould not have been offered even fiveyears ago. But it is a model that isattracting increasing approval from earthscientists as they examine the newevidence.

But what does this mean for humanstoday? Potentially a great deal,especially in understanding thegreenhouse effect and current threats tobiodiversity. The growing research linksbetween palaeontological andgeological studies of mass extinctions ofthe past, and the current ecological crisison the Earth are explored in the lastchapter. Scientists, conservationists and

politicians are looking ever moreclosely at the evidence from the past. Inplanning for the future, perhaps the bestevidence comes from the past.

Since 2003 …

Inevitably, many new discoveries havebeen made in the five years since thisbook was first published. I havecorrected minor errors in the main text,and update some topics here – many ofthe changes may seem subtle. Forexample, I use the date of 251 millionyears for the end-Permian event, and yetthere is debate about whether that is thecorrect date, or whether the age is more

properly 253 million years ago. Theevidence for a proliferation of fungiafter the crisis (Chapter 9 – ‘Plant die-back and catastrophic erosion’) has beenchallenged by palaeobotanists whobelieve the fungi are in fact algae. Wehave carried out further fieldwork on thePermo-Triassic boundary in Russia(Chapter 10 – ‘The final solution’) in2004 and 2006, and this has shown thatthe extinction of amphibians and reptileswas indeed sudden, but also that theecological recovery after the eventlasted for at least 20 million years.

Further evidence has emerged since2003 that confirms the end-Capitanianand end-Olenekian extinctions that pre-and post-date the Permian-Triassic

boundary event (Chapter 11 – ‘HowMany Events?’). Luann Becker andcolleagues published new evidence in2004 for impact (Chapter 11 –‘Buckyballs and impact’) – a crater,suevite and shocked quartz, from deepboreholes off northwest Australia – butthat evidence has been hotly debated bymost colleagues.

The suggestion that the release ofmethane hydrates is the key to the earth-bound extinction model (Chapter 11 –‘The methane burp’) has beenchallenged by evidence that there werethree or four major negative carbonisotope shifts through the 5 million yearsof the Early Triassic: the argument isthat the methane hydrates could explain

one isotope shift, but the deep methanereservoirs could not have recharged fastenough to drive repeated shifts. Thesource of those later negative carbonshifts, though, is still uncertain.

Does extinction matter?

It is easy to take highly polarized viewson the current ecological situation. Onthe one hand, the doom-mongers declarethat everything is lost and that life willbecome entirely extinct within a fewhundred years. Their calculationssuggest that we are losing 2000 speciesper day, and that our every advance incivilization is murderous. Humans

should instantly cut their populationsizes and return to self-subsistence,stone-age modes of life.

The opposing view looks at the pasthistory of human population growth andto agricultural advances, and assumescomplacently that nothing has reallygone wrong. Who really cares about theodd dodo, panda or elephant? Clearly, insome way, it is natural to expect thatsuch an advanced species as Homosapiens will take over the world, anduse its great brain to discover ever moreefficient ways of producing food andcuring disease.

Such standpoints are too extreme formost people. It is clear that humanpopulations cannot be cut back to pre-

industrial levels, nor would manypeople choose to live in rude huts andsubsist on brown rice. Equally, theworld was not created simply for humanbenefit. This idea is somewhat akin tothe creationist viewpoint that horseshave conveniently curved backs so thatthey may carry human riders. Thecomplacent assumption that ever morefood can be squeezed out of a limitedarea of agricultural land around theworld is absurd. Advancing populationsand farming are daily destroying vasttracts of habitat. But how can one draw arealistic line between the doom-mongersand the growth-enthusiasts?

Look to the past. Extinctions havehappened before, and life has recovered

in different ways. Comparison of presentcrises with documented ancientexamples at least allows scientists andpolicy-makers to work with real factsand figures. Questions can be posed: ‘If10% of species are wiped outworldwide, what happens?’, ‘If half thespecies on one continent are destroyed,do the remainder repopulate the area, oris the continent invaded fromelsewhere?’, ‘If half of life disappears,how long does the recovery phase last?’.The step-by-step dissection of ancientmass extinctions can provide some of theanswers.

The geological time-scale. Massextinctions and minor extinctionsare shown by arrows. The ‘big five’are shown with heavier arrows.

1

ANTEDILUVIAN SAURIA

On 5 March 1845, Professor RichardOwen wrote an account of some bonesof ancient reptiles that had been broughtback by Sir Roderick Impey Murchisonfrom his long peregrinations in Russia in1840 and 1841. Owen determined onesection of a backbone as

belonging to the Crocodilian division of Sauriaby the strong, short, rib-like processes from

the sides of the two anchylosed sacralvertebrae,— a modification … introduced togive a firm ‘point d’appui’ to the hinderextremities of those higher Sauria whichoccasionally walk on dry land.1

Owen then went on to describe otherbones he had been sent. He comparedthem all to modern crocodiles, and to areptile called Thecodontosaurus whichhad been named in 1836 from theTriassic rocks of Bristol in southwestEngland.

What does this passage mean? Allthat Owen was in fact saying was thatthe sacral vertebrae, that is the elementsof the backbone that lie in the hip region,showed strong ribs at the side to whichthe hip bones could attach. Such a strong

attachment is seen today in crocodilesand other reptiles that are well adaptedfor walking on land. He was stressingthe fact that this was not some primitivekind of animal that lived most of the timein water and hence would have no needfor a strong skeletal framework for thelegs. An advanced reptile such as theRussian crocodile indicates that it comesfrom rocks of Triassic age, or later.

Owen’s determination of the Russianreptiles, arcane as his language may be,was critical evidence used by RoderickMurchison in his interpretation of thePermian System of rocks in Russia.Murchison had just invented the name‘Permian’ for the fossiliferous sedimentsaround the city of Perm on the west side

of the Ural Mountains, and, byimplication, for all rocks of the same agefrom other parts of the world. Inaddition, for Owen, these specimensoffered further information towards afuller understanding of the early historyof the reptiles.

It is important to place Owen’sviews in context, and examine wherethey came from. In the absence ofadequate specimens, and with extensiveconfusion between beasts from what wenow recognize as Permian and Triassicrocks, it is no wonder that savants ofearly Victorian times could not conceiveof a vast end-Permian extinction event.Let us look briefly at what we know nowof amphibian and reptile evolution, and

then piece together what Owen knew in1845.

Tetrapods: batrachian andsaurian

Modern amphibians and reptiles areeasy to tell apart.2 The amphibians, suchas frogs and salamanders, generally laytheir eggs in water, and from these hatchtadpoles. The tadpoles develop asentirely aquatic creatures – essentiallyas fishes – before they metamorphoseinto the adult form which lives partly onland and still partly in the water. Indeed,the name amphibian means ‘life on both

sides’ (from the Greek amphi, ‘on bothsides’, and bios, ‘life’), in other words,life both in the water and on land.

Modern reptiles, particularly lizardsand snakes, live their lives entirely onthe land and are adapted to life in dryconditions, with horny waterproof scalesover their bodies. They lay eggs withwhite calcareous shells (like a hen’segg) or with a leathery coat – in eithercase, the egg is waterproof and theyoung reptile develops safely insidebefore hatching straight into life on land.

Naturalists have long realized thatamphibians are essentially intermediatebetween fishes and reptiles. Today, ofcourse, we see this as an expression ofthe evolution of the respective groups –

fishes came first, then amphibiansevolved from fishes when they took theirfirst faltering steps on to the land, andfinally the reptiles evolved from theamphibians by dispensing with life in thewater altogether. Owen was famouslynot an evolutionist – he was later one ofCharles Darwin’s most forceful critics –although he could not deny the apparentsequence of forms.

But, in the 1840s, the fossils wererare. Equally, the dividing line betweenthe fossil amphibians and reptiles washard to draw. Owen, and others at thetime, frequently referred all the earlytetrapods – that is, four-footed creatures– to Reptilia. The older, amphibian-likeones were sometimes called Batrachia,

or Batrachian Reptilia, and those thatcould be compared more closely withmodern lizards or crocodiles weresometimes called Sauria.

Tetrapods of the coal forestsand Permian deserts

We now know that the first tetrapodscame ashore in the Late Devonian, some370 million years ago, and that thesebasal tetrapods – amphibians – radiatedinto a variety of forms in the subsequentCarboniferous period. TheCarboniferous amphibians includedland-living forms with four sturdy limbs,

as well as many largely aquatic species,some of which swam with their paddle-like limbs, and others of which had losttheir limbs entirely. They ranged in sizefrom just a few centimetres to 3- and 4-metre long monsters.

The Carboniferous is famous for itscoal forests – great tangled masses ofvegetation growing luxuriantly in thetropical conditions that wereexperienced all over Europe and NorthAmerica at the time (both continentsstraddled the Equator during theCarboniferous). Crawling through thelow vegetation were myriad millipedes,centipedes, spiders and the like. Higherin the trees were insects, some of themremarkable for their huge size – there

were dragonflies as large as small gullsin those days. The amphibians feasted.

Among the Carboniferousamphibians were a few rather moreterrestrialized forms, animals that hadwaterproof skins and which laid eggswith shells. At first, these basal reptilesdid not make much of a mark, since theamphibians, with food and damp habitatsin abundance, flourished. But the greattropical forests of Europe and NorthAmerica began to dry out towards theend of the Carboniferous and into thesubsequent Permian period. Most of theamphibians died out, and only somesmaller aquatic forms survived in thereduced watercourses.

Now it was the turn of the basal

reptiles to flourish. During the Permian,the group called the synapsids, ormammal-like reptiles, rose todominance, especially in the LatePermian. Ecosystems became complex,with synapsids of all shapes and sizesfeeding on the plants, insects and otherinvertebrates, and each other. By the endof the Permian, some ecosystems wereas complex as any today. Top-levelanimals had evolved – massiveherbivores the size of rhinos, and sabre-toothed carnivores that could pierce thethick hides of these huge plant-eaters.Among the smaller reptiles, in additionto the synapsids, were alsorepresentatives of other groups,including forms distantly ancestral to

turtles, and to crocodiles and lizards.But these animals abruptly

disappeared, wiped out by themysterious end-Permian crisis. Thecomplex ecosystems collapsed and allthe richness of Late Permian life wasdestroyed. What came next was one ofthe most extraordinary times in thehistory of the Earth.

When pigs ruled the Earth

A few years ago, I had a rather plaintivephone call from a researcher whoworked for a small, independent film-maker near London. They had completed

a one-hour documentary about the end-Permian mass extinction, but theprogramme had just been turned downby the commissioning televisioncompany. I asked why, and was told itwas because there was too much aboutlemurs in the film. This surprised me,since lemurs have absolutely nothing todo with the end-Permian mass extinction– indeed, the first lemur fossils dateback only a million years or so, and theirdistant relatives, some of the firstprimates, are at most 60 million yearsold.

It turned out that the film-makers hadspent all their budget flying toMadagascar and filming the rare lemursand other primitive primates of that

island. I asked them how much footagethey had of Permo-Triassic rocks inRussia and South Africa – the sequencesthat span the geological boundary andthat contain clues that might point to thecauses of the event. I detected a gulp atthe other end of the line. The researchershad clearly failed in their job, and thefilm-makers had panicked and thought,‘if we can’t fill the hour with dinosaurs,let’s go for the monkeys’. They did,however, have animatronic rubberdicynodonts.

The dicynodonts were one of thosekey synapsid groups that had been hugelydominant in the Late Permian. Frompoles to equator, the dicynodonts werethe major plant-eaters, some of them

reaching the size of hippos. It was wellknown that they had been virtuallywiped out by the end-Permian massextinction, and only one form hadsurvived: a medium-sized dicynodontcalled Lystrosaurus. Subsequently, fromLystrosaurus, the dicynodonts re-flowered in the Triassic, radiating backto something like their former diversity.

The film-makers had at least heardo f Lystrosaurus, this extraordinaryreptile that had survived the end-Permian cataclysms. All the impacts,eruptions, freezing, poisoning and othercatastrophic events had not discouragedLystrosaurus, which not only survived,but repopulated the Earth fromAntarctica to China, from Argentina to

Russia. Surely Lystrosaurus must havebeen the original bionic organism,possessing superpowers that its friendsand relatives lacked?

To rescue their film, I wasinterviewed at the animatronics studio inLondon. As the model Lystrosaurusgulped and rolled its eyes beside me, Itried to explain that it was in fact a veryordinary animal. It had no specialsurvival qualities that the other animalslacked. It was simply lucky. ‘Why’, theykept asking, ‘did Lystrosaurus survive,and nothing else?’ A clear adaptivestatement was required. I explained thatLystrosaurus was a survivor of a sort,but it was not particularly fast, fearsomeor intelligent. It was really something

like a Triassic pig in appearance, andeven perhaps in habits, since it wasprobably a generalist, without specificadaptations in its diet, livingrequirements or mode of locomotion.

The point is that good fortune is acharacteristic of mass extinctions. Thesurvivors are more lucky than speciallyadapted. The most advanced, intelligent,fast-breeding animal species may bewiped out by the chance calamity of anextinction event when it is obliged toface challenges that have never beenencountered before. Evolution works tohone the fine details of the adaptations oforganisms against commonlyencountered problems, such as droughts,floods, predators and diseases, but rare

events that happen perhaps once everyfew million years just cannot beaccommodated. The phenomenon is whatthe palaeontologist David Raupmemorably described as ‘bad luck, notbad genes’.

S o , Lystrosaurus was not aspecially tough survivor, a perfectanimal to weather the storms and found amajor new dynasty. It was rathernondescript, about 1.5 metres long, witha bulky, blimp-like body and ratherinadequate hindquarters (Fig. 1). Its legswere short and its head was heavy,armed with a small snout and horn-edged jaws that lacked all but the canineteeth. Lystrosaurus was a plant-eaterthat evidently sliced tough stems and

chopped them into manageable fragmentsby a circular backwards-and-forwardsrotatory jaw action. All of this wasnicely shown by the animatronic modelwhich, by the operation of a controlpanel, could be made to snuffle or drool,while technicians added suitable snortsand howls.

1 Lystrosaurus, survivor of theend-Permian mass extinction.

A few weeks later, when the filmhad been put together, it was re-titled‘When pigs ruled the Earth’, the titleappearing in all the pre-publicityhandouts and notices. The Sunday Timesran a full-page story, accompanied bylurid colour illustrations, about how pigshad been hugely successful in theTriassic, and how they had later evolvedinto the array of other mammals weknow and love today. On the day afterthe film was shown, I encounteredwithering looks of pity from colleaguesand students alike. But at leastLystrosaurus became known to a wideraudience.

Reptilian renaissance in theTriassic

The earliest Triassic was indeed ableak, impoverished time. Gone werethe specialized plant- and flesh-eaters ofthe Late Permian. All that remainedwere literally one or two tetrapodspecies worldwide, includingLystrosaurus, and these slowlypopulated the Earth, eventually givingrise to a second flowering of thereptiles. But the old days were over.New animals came on the scene,including the dinosaurs. Some synapsidsre-radiated in the early part of theTriassic, and they diversified to a

certain extent. But ecosystems did notreach the levels of complexity achievedin the Late Permian for millions of years.And the synapsids were unable, for along time, to regain their formerascendancy.

In fact, the end-Permian massextinction triggered an evolutionarydance. The synapsids largely withdrew,and their place was taken by the greatreptilian group called the diapsids (‘twoarches’, referring to their two cheekbones behind the eye socket),represented today by crocodiles, lizards,snakes and birds, and deriving from adistant, Carboniferous ancestor.

During the Carboniferous andPermian, the diapsids had been a minor

element of most faunas, only a smalllizard-like creature here and there, neverlarge, and rarely more than 2 or 3% ofthe total numbers of animals. Someadmittedly took to gliding in the latestPermian, but they were hardly a majorgroup. In the Early Triassic somediapsids, in particular the group calledthe archosaurs (‘ruling reptiles’) tookover the carnivore niches. They preyedon the re-evolving synapsid plant-eaters.During the first 20 million years of theTriassic, the basal archosaursdiversified slowly, and eventuallyincluded huge predators, some of themup to 5 metres long. Then, in the LateTriassic, big changes occurred. The firstdinosaurs appeared – rather small,

bipedal forms at first, they diversifiedrapidly and soon reached huge size.Mammals were around during the entireage of dinosaurs, small descendants ofthe formerly dominant synapsids, but thedinosaurs ruled the Earth for the next165 million years, until their extinction65 million years ago. Then, and after along wait in the wings, the synapsiddescendants, the mammals, finallymoved back to dominate the Earth, aposition they had last held in the LatePermian. The synapsid-diapsid-synapsidcycle had gone full-circle.

This is our present understanding,founded on extensive collecting andstudy of specimens from around theworld since the 1840s. But Richard

Owen knew very little of this. He hadinformation from only scattered fossilremains, and indeed, many of theprinciples of his work had only just beenestablished. In particular, the newscience of comparative anatomy, whichwas Owen’s forte, was only around 20years old when he was studying theRussian saurian fossils. No wonder thatneither he, nor any of his colleagues,could even conceive of the magnitude ofthe mass extinction event that they hadjust begun to document.

The nature of fossils

I remember collecting my first fossils at

the age of eight. I was brought up inAberdeen, in northeast Scotland, a citywhere all the older buildings areconstructed from the native silver-coloured granite, a resilient rock thatlooks as fresh today as it did when itwas cut 100 or 200 years ago. Granite isan igneous rock, formed deep in theEarth’s crust from molten magma, andthere was no chance of ever findingfossils in that.

On family holidays in Yorkshire,though, I was able to explore morefruitful rocks, in particular the marineJurassic rocks of the coast aroundSaltburn, Staithes and Whitby. There itis easy to collect fossils – clams,brachiopods, coiled ammonites, rare

corals and even isolated bones of marinereptiles and fishes. Easiest to find werespecimens of Gryphaea, formerly called‘devil’s toenails’ since they look like thecalloused, horny toenails of some greatdragon. A Gryphaea specimen fitsneatly into the palm of a child’s hand – adeeply curved shell, shaped somethinglike half a doughnut, but with a flat shellcovering the top. Gryphaea, an oyster,lived with its curved lower shell half-buried in the sea-bed mud, and its upper,flat shell able to flap open and shut,rather like the lid of a kitchen bin. Icollected 160 of these shells on my firstfossil-hunting trip, and insisted that wecarry every one of them back with us toour hotel, and then on the train home to

Aberdeen.It was obvious to me then that these

fossils were the remains of animals thathad once been alive. Of course, todaywe read that fact in all the books aboutfossils, and any child who is interestedin dinosaurs will know that the shells,bones and impressions buried in therocks can be brought back to life bypalaeontologists. But this has not alwaysbeen self-evident.

In the Middle Ages, scientistsdiscussed the nature of fossils long andhard. Surely, they argued, the fossilshells that they found high in the hills ofItaly or France could not be the remainsof real sea creatures, since the sea wasmiles away? Some of them might have

been carried there by Roman soldiers,but not all of them. As in so many things,however, Leonardo da Vinci had thecorrect answer. He argued that if a fossillooks in every detail like a modern shellor bone, then it must clearly be theremains of some ancient organismrelated to the modern forms. The fossilshells got into the high hills since theseas had once covered those hills, or thehills had been lifted up after the rockshad been deposited, to form themountains we see today.

The debate about the nature offossils continued actively through theseventeenth century. The great savants inEngland and France presented weightyarguments on both sides. Some claimed

that the fossil shells, sea urchins, sharks’teeth, mammoth bones and other fossilsthey found were indeed the remains ofancient organisms. Others argued thatthis was impossible: surely the so-calledfossils had been placed in the rocks bythe Almighty to test our faith. Sincefishes and molluscs clearly could notlive in rocks, the fossils were no morethan inorganic productions, the result ofthe action of plastic forces deep withinthe Earth. To lend philosophicalseriousness to their proposals, theopponents of fossils as the remains ofancient plants and animals often used theterm vis plastica for ‘plastic force’, theLatin version presumably sounding moreauthoritative.

But as the fossils piled up, thismiraculous view had to be rejected.However, if all the fossils representedplants and animals from long ago, thisopened up two further difficultquestions: first, where had all theseplants and animals gone, and second,how long ago had they lived?

The Yorkshire crocodile andthe Ohio incognitum

By 1750, most naturalists accepted theorganic nature of fossils, but they stilldenied the possibility of extinction. Thismay seem paradoxical, but any Christian

of the time could not readily accept thisnotion, since extinction, the loss of oneor more species, would imply an errorof some kind in God’s plan. But if theywere not extinct, where now were allthese strange plants and animals of thepast as represented by the fossils, but notknown to be alive anywhere?

One way out of the dilemma was tosuggest either that the fossil creatureshad lived before the biblical GreatFlood, or that they perhaps still existedin some unexplored part of the world,such as the far reaches of NorthAmerica, Australia, India, or any of theother lands that were just then beingexplored by Europeans.

In 1758, a beautiful fossil of a

crocodile was reported from the foot ofa cliff on the Yorkshire coast, nearWhitby – indeed close to the site of myearly fossil-hunting forays, as it happens.The crocodile skeleton, laid out as blackfossil bones, all in their correctrelations, was embedded in Liaslimestones, recognized now as part ofthe Jurassic. The specimen wasdescribed in two separate reports to thePhilosophical Transactions of theRoyal Society of London, by WilliamChapman and a Mr Wooler, whose firstname has been lost to history. Woolerargued that the fossil was identical ‘inevery respect’ to the modern Indiancrocodile from the River Ganges. As toits position, at the foot of the cliff and

buried under 180 feet (55 metres) ofrock, there could be no doubt3

that the animal itself must have beenantediluvian, and that it could not have beenburied or brought there any otherwise than bythe force of waters of the universal deluge. Thedifferent strata above this skeleton never couldhave been broken through at any time, in orderto bury it to so great a depth as 180 feet; andconsequently it must have been lodged there, ifnot before, at least at the time when thosestrata were formed.

This was all very well for a fossilanimal that apparently looked like aliving form, but what about somethingthat seemed to have no obvious livingcounterparts?

Such was the mastodon.4 The first

specimens were found on the banks ofthe Ohio River in 1739 by a FrenchCanadian army officer, Baron Charlesde Longueuil, who sent the bones andteeth back to Paris. There, a great furorebroke out, as the leading naturalists ofthe day debated whether the mastodoncould be extinct or not. Furtherspecimens were collected by Britishsettlers, and they were discussed inLondon, but, by 1769, the issue waspretty much settled.

The intervention of the greatanatomist and physician, William Hunter(1718–83), was critical. He comparedthe jaw bone of the American mastodon,which he termed the Ohio incognitum,with that of a modern elephant, and saw

that they were identical, except for theteeth. So this must have been anelephant, but one that was not the sameas either the African or Indian elephantsof today. Hunter pursued his argumentremorselessly. He had examined most ofthe specimens sent both to Paris and toLondon, and he therefore concluded withthe comment that, ‘though we may asphilosophers regret it, as men we cannotbut thank heaven that its wholegeneration is probably extinct’.

Georges Cuvier (1769–1832), thegreat French anatomist and geologist,5finally resolved the extinction issue,with a flurry of publications in 1795, atthe beginning of his glittering career inParis. He wrote accounts of the

mastodon and of the Siberian mammoth,many specimens of which had been dugout of the frozen tundra, some evenretaining hair and flesh. His third casestudy was the giant ground slothMegatherium, a huge animal that livedformerly in the Americas, particularlySouth America.

Cuvier used a new scientificapproach in these papers, one that cameto be known as comparative anatomy. Inthis approach, the anatomist comparesevery fine detail of the skeleton andother structures of the body of differentanimals with a view to findingregularities and similarities of structure.By using his thorough knowledge of theanatomy of modern animals, Cuvier was

able to demonstrate that these ancientanimals were indeed similar to modernforms – to elephants and tree sloths, forexample – but they were clearly alsoanatomically distinct. These were alllarge mammals, and unlikely to bemissed if they were still alive, Cuvierargued, and his case was so ablydemonstrated that the last doubters hadto accept the reality of extinction.

This is ancient history now, but toRichard Owen the debates were justsettling down. As a young man studyingmedicine and anatomy in Edinburgh inthe 1820s, Owen heard constantreference to the great Cuvier. Britishanatomists looked to Paris for the best,most up-to-date thinking, and stories of

Cuvier’s abilities spread. It was said,for example, that he could identify anyanimal from a single bone, and hedemonstrated his skill in publicpresentations of his discoveries. Owenset himself the task early in his life to tryto emulate the great Frenchman. But thatwas a long time in the future. In 1845, ashe pondered the unusual new reptilebones from the Russian Permian, Owenrecalled some obscure references tosuch beasts having been found in theprevious century, at the height of themastodon-mammoth-extinction debates.

First finds in the Russian

Copper Sandstones

The first finds of the famous Permianfossil reptiles from Russia came to lightin the 1770s, but only in obscure ways.6The fossil bones were found in theCopper Sandstones, a belt of ore-richrocks stretching for hundreds ofkilometres along the western slope of theUral Mountains. The specimens camefrom the extensive mining works forcopper, especially from a small numberof mines in what is now OrenburgProvince, as well as in Bashkortostanand some from Perm Province. TheRussian Academy of Sciences sent outscientific expeditions to the UralMountains from 1765 to 1805, but the

naturalists did not always understandwhat they had found. One of them, P. I.Rychkov, mentioned the discovery offossil reptile bones in the CopperSandstones in his diary of 1770, but hetook them to be the remains of ancientminers.

The Copper Sandstones reptilefossils continued to be collectedsporadically through the late eighteenthcentury, mainly by mining engineers whohad a personal interest in these curiousremains. As the mines were so remotefrom fashionable St Petersburg,however, little was reported to thescientific societies there and hardly anyformal study was undertaken. Distancesin Russia are huge, and in those days the

roads were appalling, so the fastestpost-chaise would take at least fourweeks to cover the 2500 kilometres fromPerm or Ekaterinberg back to StPetersburg. To the French-speaking tsarsand princes in their glittering court, themineral workings of the Ural Mountainsmight as well have been on anotherplanet. And almost nothing of theseremarkable discoveries filtered out tothe West until somewhat later. Cuvierhad nothing to say about the RussianPermian reptiles because he had neverheard of them, and he had died beforeOwen received some of the firstspecimens to reach the West, fromMurchison. But Cuvier did play a largepart in the interpretation of the first

dinosaurs to be found and identified.

Giant saurians

Bones of what we now know to havebeen a large meat-eating dinosaur7 cameto light in Middle Jurassic rocks north ofOxford in about 1818, and were taken toWilliam Buckland (1784–1856),Professor of Geology at the Universityof Oxford and Dean of Christ Church. Atthat time it was not uncommon forchurchmen to combine their careers withscience. Buckland was a brilliant mix ofphilosopher and eccentric. His housewas full of fossils and he was sought

after for his impressive lectures toundergraduates about geology and thelife of the past. His dinner partyinvitations were received with a mixtureof pleasure and nervousness: pleasure atthe thought of his agreeable company andamusing conversation, but nervousnessat what might grace the table. Bucklandwas famous for having eaten his waythrough the entire animal kingdom. Heenjoyed most of his culinaryexperiments, but could not find anagreeable way to prepare house flies.

Buckland could not identify the hugefossil bones, and he showed them toCuvier in Paris and to other experts. Inthe end, Buckland classified the animalas a giant reptile, probably a lizard, and

he estimated it had been 40 feet (12metres) long in life. After six years ofconsideration, Buckland finallypublished a description of the bones in1824, and he stated that they came froma giant reptile which he namedMegalosaurus (‘big reptile’). This wasthe first dinosaur to be describedformally.

At the same time, independently,Gideon Mantell (1790–1852), a countryphysician in Sussex, was amassing largecollections of fossils. Mantell was asfanatical about geology andpalaeontology as Buckland, but helacked the family background and wealthto indulge his philosophical pursuitsfreely. His life was dogged by the

tension between his passion for fossilsand the need to earn a living and toachieve some kind of position in society.During a visit to a patient near Cuckfieldin 1820 or 1821, so the story goes, hiswife Mary, who had come with him,picked up some large teeth from a pile ofroad-builders’ rubble. The teeth hadevidently come from the Wealden beds,recognized now as Early Cretaceous inage.

Mantell realized the teeth belongedto some large plant-eating animal, but ashe was not skilled in anatomy he sent theteeth to Cuvier, who assured him that theanimal must have been a rhinoceros.Mantell then compared the teeth withthose of other modern animals in the

Hunterian Museum in London, and astudent there, Samuel Stutchbury,showed him that they were like the teethof a modern plant-eating lizard, theiguana, except they were much bigger.So, Mantell described the seconddinosaur, named by him Iguanodon(‘iguana tooth’) in 1825, based both onthe teeth and on some other bones he hadfound since.

William Buckland and GideonMantell had great difficulty ininterpreting these early dinosaur bonessince there were no modern animals tocompare them with. In the end, theydecided that Megalosaurus andIguanodon were giant lizards. The thirddinosaur to be named, Hylaeosaurus,

came to light in southern England in thesame Wealden beds that had producedIguanodon. Mantell named it in 1833.Further dinosaurs were named in the1830s, including two from the Triassic:Thecodontosaurus was named in 1836by Henry Riley and Samuel Stutchburyfrom Bristol, southwest England, andPlateosaurus in 1837 by Hermann vonMeyer from southern Germany. In 1842,Richard Owen himself named the sixthdinosaur, a giant plant-eater, again fromthe Middle Jurassic of Oxfordshire,Cetiosaurus. But still no one knew whatthese extraordinary giant beasts actuallywere.

Richard Owen, the ‘BritishCuvier’

By 1841, when Murchison brought backhis valuable bones from the Permian ofRussia, Richard Owen (1804–92) wasrising rapidly in reputation.8 Owen hadarrived in London in 1825 after hismedical training in Edinburgh, and wasappointed as an assistant at the RoyalCollege of Surgeons in 1826, with thetask of sorting out their collections ofanatomical specimens, both human andanimal. He began to produce a series ofcatalogues in 1830, continuing with thetask until 1860. In 1836 Owen wasappointed Hunterian Professor of

Anatomy at the Royal College ofSurgeons. A reward for his labours, thisalso gave him the status he craved inorder to make an impact in scientificcircles in London.

Although he had worked mainly onthe anatomy of living (or at least recentlydead) animals up to 1836, his new statusas the ‘British Cuvier’ made him theobvious choice to review the burgeoningmaterials relating to the giant fossilreptiles. Owen at the time was aware ofhis own intelligence, and he wasintensely ambitious. With his highforehead and arresting eyes (Fig. 2), hiscapacity for fast and accurate work andhis skills as a lecturer, he made a strongimpression on his older colleagues.

Thus Owen, a scientific gentleman ofthe new breed – one who earned hisliving – was awarded one of the firstscientific research grants forpalaeontology. In 1838, the BritishAssociation for the Advancement ofScience (the BA) offered him £200 topay his expenses in compiling a surveyof the British fossil reptiles. For this hetravelled extensively throughoutEngland, examining specimens in publicmuseums, but many more in thecollections of wealthy amateurs. Hepresented his report on the marinereptiles, the ichthyosaurs andplesiosaurs from the Jurassic rocks ofthe Dorset and Yorkshire coasts, andfrom the Cretaceous of Kent, at the BA

meeting in 1839 in Birmingham.

2 Portrait of Richard Owen, theman who named the Dinosauria, ashe was around 1840.

The committee of the BA was sopleased with Owen’s work that theypromptly awarded him a further £200 to

extend his survey to the terrestrialreptiles – the crocodiles, turtles andgiant saurians. Owen set off again on histours around England, travelling mostlyby carriage but using the new railwayswhere he could. He prepared a full anddetailed account of all that he had seenduring the course of two years of intensework, and he presented his paper at thenext meeting of the BA, this time inPlymouth, on 2 August 1841. The talklasted for two and a half hours, and theaudience was reported in contemporarynewspaper accounts to have received theperoration with satisfaction andapprobation. In any case, much pleasedwith his brilliant overview, the BAawarded Owen a further sum of £250 to

complete another report, and £250towards the cost of engravings of theextinct reptiles.

Owen’s work was straightforwardas regards the fossil turtles andcrocodiles: they could be comparedreadily with their living relatives.However, the giant saurians –Megalosaurus, Iguanodon,Hylaeosaurus and the rest – were muchharder to explain. In the first version ofhis report, the basis for the talk inPlymouth, Owen simply tried to shoe-horn them all into existing groups. Sosome were overblown lizards, otherswere crocodiles, and yet others weresomething in between. But Owen wasmuch more daring in the final report,

which was published some months later,in 1842.

Owen names the Dinosauria

Between August 1841 and April 1842,Owen had a revelation. No longer werethe giant saurians to be explained assome ancient and overgrown tribe oflizards. They clearly represented anentirely distinct group that had left nodescendants. Owen’s revelation came ashe studied additional specimens ofIguanodon and Megalosaurus. He notedthat the femur (thigh bone) looked moremammalian than reptilian in certainfeatures. In particular, the head of the

femur was curved inwards and had aball-like structure that would fit into asocket made by the hip bones. Thesegiant saurians apparently stood high ontheir legs, just as cows, humans andbirds do; they did not sprawl, like livinglizards and crocodiles.

Owen then saw a new specimen ofIguanodon that had come from the Isleof Wight, which preserved the hip regionof the backbone. The sacral vertebrae,the elements of the backbone to whichthe hip bones are attached, weredifferent from those of any living reptile.Normally, reptiles have two sacralvertebrae, and they are separate anddistinctive bones. The new specimenshowed that Iguanodon had five sacral

vertebrae, and they were firmly fusedtogether, and to the hip bones. Owenrecalled that he had seen the samearrangement in a specimen ofMegalosaurus in Oxford.

S o Megalosaurus, Iguanodon andHylaeosaurus differed from modernreptiles by being huge, by having amammal-like femur and by their fusedsacral vertebrae. Owen inserted a fewbrief paragraphs into the manuscript ofhis BA talk, and the deed was done. Thegiant land reptiles from the BritishJurassic and Cretaceous were not hugelizards, but were representatives of anew group. He wrote:9

The combination of such characters, some, as

the sacral ones, altogether peculiar amongReptiles, others borrowed, as it were, fromgroups now distinct from each other, and allmanifested by creatures far surpassing in sizethe largest of existing reptiles, will, it ispresumed, be deemed sufficient ground forestablishing a distinct tribe or suborder ofSaurian Reptiles for which I propose the nameof ‘Dinosauria’.

In a footnote, Owen explained his newterm, based on the Greek words deinos,meaning ‘fearfully great’ and saurosmeaning ‘lizard’, or in more modernterms, ‘reptile’. So Owen was father ofthe dinosaur, and in naming the group hedisplayed his synthetic powers and hisbrilliance. Gideon Mantell, who hadlaboured for twenty years on the giantbones, had failed to see how his new

fossils fitted into the broader picture ofthe history of life, and he was rapidlyeclipsed by his younger, more urbanerival.

Owen’s new understanding of thedinosaurs of the Jurassic and Cretaceousallowed him to determine that the bonesshown to him by Murchison in 1845were clearly not from a dinosaur.Indeed, he argued that they came from areptile that was more ancient than adinosaur, and this was a criticalobservation for understanding the placein the overall rock sequence of the newRussian rocks that Murchison had calledthe Permian.

Owen and Murchison’sRussian reptiles

The Copper Sandstones of the Orenburgdistrict in the Urals had, as we haveseen, been a source of the bones of fossilreptiles10 since 1770. But little seriousnotice was taken of the fossils until the1830s. In 1838, S. S. Kutorga, aprofessor at the University of StPetersburg, gave the first scientificdescriptions of the fossils, naming oneBrithopus, on the basis of somefragments of the humerus (the upper armbone), and the other Syodon, on the basisof a tusk fragment. Kutorga made theastonishing claim that these were early

mammals – Brithopus a relative of themodern sloths and anteaters, and Syodonof the elephant. In retrospect, this seemsbizarre, since we know now that the firstsloths and elephants appeared only some50 million years ago, not 260 millionyears ago in the Permian, before the ageof the dinosaurs.

Kutorga’s claim was not soremarkable in another way, however. Inspotting some mammalian characters inthe Copper Sandstone fossils, he hadactually unwittingly hinted at what theseextraordinary reptiles might be,something that Owen was not to realizefor a long time. These were indeedmembers of the synapsid line of reptiles,and therefore nothing to do with frogs,

dinosaurs, lizards or any of the otherfossil groups that Owen was thenstudying. The Permo-Triassic synapsidsare commonly known as mammal-likereptiles, to signify their evolutionaryposition on the way to mammals.

Independently, one of the directorsof the copper mines in Ufa (now inBashkortostan Province) and Orenburg,the splendidly named F. Wangenheimvon Qualen, had amassed one of thelargest collections of Copper Sandstonereptiles. As his name suggests, vonQualen was of German extraction, not souncommon in Russia at that time, whenmany scientists and technical expertswere imported from Prussia. VonQualen published a series of reports in

the Russian scientific journals on thebones he had found. He also speculatedabout the age of the reptile beds, and hecorrelated them with the Kupferschiefer(‘copper shales’), a part of the Zechsteinof Germany. As a mining engineer, vonQualen was familiar with theKupferschiefer, a rock unit that had beenmined for decades in Prussia. And hisage correlation turned out to be inspired:the Zechstein is indeed now dated asLate Permian in age, as we shall see inChapter 2.

Von Qualen’s specimens werestudied also by Kutorga, and by twoother noted professors, German importslike himself: J. G. F. Fischer vonWaldheim, from the Natural History

Museum in Moscow, and E. I. vonEichwald, from the Academy of Sciencein St Petersburg. Waldheim named anew mammal-like reptile, Rhopalodon,and Eichwald named the reptileDeuterosaurus and the amphibianZygosaurus.

Owen did not have a chance to seethese Russian specimens at the time,though he knew of the publications, andin reporting on the bones brought backfor him by Murchison, he identified themas belonging to Fischer’s beastRhopalodon. However, as we saw at thestart of this chapter, Owen believed thatthis reptile was a crocodile of somekind. Whether it was a crocodile or amammal-like reptile, Owen’s studies

confirmed for Murchison that theRussian sandstones of the south Uralswere part of a new system of rockswhich he called the Permian. Thereptiles seemed to be most like those ofthe Triassic of England and Germany,and that was all he wanted to hear.

Revolution in palaeontology

In a few short decades, then,palaeontologists had resolved somemajor debates. Extinction was real, asCuvier had convincingly shown in the1790s, and fossils could be studied asthe remains of plants and animals thathad once lived. Owen and his

contemporaries accepted thatsimilarities among fossils indicatedsimilarities in the ages of rocks indifferent parts of the world. Cuvier’snew science of comparative anatomygave Owen the tools to make thesecomparisons, and to begin to try todisentangle just what had been going onin the history of life through the Permian,Triassic, Jurassic and Cretaceous. Inretrospect, some of his ideas may seemodd, but he was limited by the isolatedfossil finds and the difficulties oftravelling from place to place to makenecessary comparisons.

So much for Owen’s understandingof palaeontology and comparativeanatomy. The other great revolutions in

natural science in the early nineteenthcentury had turned geology on its head.These rapid advances allowedMurchison to identify rocks of a newgeological system, the Permian, and thatwas a necessary precursor to our currentunderstanding of the greatest massextinction of all time.

2

MURCHISON NAMES THEPERMIAN

In December 1841, RoderickMurchison,1 then President of theGeological Society of London, namedthe Permian System, based on rocks hehad observed around the city of Perm inRussia. He did this in a short note of sixpages in the Philosophical Magazine

and Journal of Science, under the rathernon-committal title ‘First sketch of someof the principal results of a secondgeological survey of Russia.…’. It tookthe form of a letter written to M. Fischervon Waldheim, bearing the date 5November.2 This seems a modest way inwhich to make such a majorannouncement, and the publication of anapparently private letter might alsoappear odd. It also looks like it wasdone in haste, published exactly onemonth after it was sent in. What wasgoing on?

After a lengthy tour of the Urals andsouthern European Russia, Murchisonhad spent a few days in Moscow, inOctober 1841, completing various

pieces of business before his returnhome. It was here that he wrote the letterto Fischer von Waldheim, outlining hisdiscoveries of the previous months.Johann Gotthelf Friedrich Fischer vonWaldheim (1771–1853), a Germanpalaeontologist, was Director of theNatural History Museum in Moscow,and, as we saw in the last chapter, hewas also one of the Russian-Germanswho had named reptiles from theRussian Copper Sandstones.

Murchison’s express instructionswere that his report, in the form of theletter to Waldheim, should be publishedat the same time in the English, Germanand Russian journals. The letter dulyappeared in Russian in the Gorny

Zhurnal, the house journal of the MiningInstitute in St Petersburg in late 1841,and in English in the PhilosophicalJournal in December 1841 and, ratherbelatedly, in the Edinburgh NewPhilosophical Journal early in 1842.Today, such multiple publication of thesame scientific paper would be severelyfrowned upon.

This was a major contribution toworld geology. Naming a geologicalsystem is not an everyday occurrence,and by 1841 most of the geological time-scale had been divided up and labelledby various geologists, includingMurchison himself. So Murchison waswell aware of the significance of hisaction. Why, then, did he choose to issue

his article at high speed in a periodicalthat was not a major geological journal?Why, indeed, in such a brief form,merely a copy of a letter sent to a fellowscientist overseas? Was he forced intorapid publication?

The reformed fox-hunter

By 1841 Roderick Impey Murchison(1792–1871) was at the height of hispowers (Fig. 3). He came from amoderately wealthy Scottish family, andhad served in the army in the PeninsularWar and in Ireland. When the fightingended, he retired from the army, marriedan intelligent heiress, Charlotte Hugonin,

in 1815, and settled down to life as acountry gentleman. But the lack ofactivity bored him and he decided totake up scientific pursuits. Soon after hismarriage, he sold the family seat atTarradale, just outside Inverness, andmoved to London, where he was able tomix with the leading scientists of the day(although they did not generally callthemselves scientists – the word cameinto use by 1840 – but rather geologists,observers, savants or philosophers). Heattended courses in chemistry at theRoyal Institution, and, having decidedthat geology was the rising subject,joined the Geological Society of Londonin 1825. Geology appealed to his desirefor physical activity, and it allowed him

to pursue his passion for pheasantshooting and fox hunting. At this point,Murchison was comfortably well off,and he did not have to work to supporthis home and family, but he only becameseriously wealthy after his mother-in-law died in 1838, and his wife Charlotteinherited her estate.

3 Portrait of Roderick Murchisonaround 1841, the year in which henamed the Permian System.

But Murchison was no idledilettante. He engaged in lengthy fieldexcursions to difficult parts of the world,and he worked with amazing speed. Thescope of his interests was vast and, atany time, he was typically involved intwo or three major fields of research. Hewas also a great controversialist. Fromthe late 1840s, after he had finished hiswork in Russia, Murchison wasPresident of the Royal GeographicalSociety for many years, and he sent outexpeditions all round the globe. Theexplorers and cartographers named

numerous mountains and lakes after him.During the 1830s, Murchison

concentrated on stratigraphy, the study ofrock sequences and the timing of eventsin Earth history. This was an excitingtime to be a geologist. Murchison andhis colleagues were literally carving upgeological time into its major divisions,and he in particular realized that thiswas a pursuit with global ramifications.In other words, close study of the rocksof one part of the world could provideevidence for a world standard ofgeological time. The standard geologicaltime-scale is such a basic part ofgeological understanding today that it iseasy to forget that things were not soclear in the 1830s.

In 1839 Murchison had named theSilurian System,3 a unit of geologicaltime characterized by certain rocks inWales (and named after the Silures, anancient tribe of Wales). Also in 1839,Murchison, together with his sometimeenemy and sometime friend AdamSedgwick (1785–1873), Professor ofGeology at the University of Cambridge,named the Devonian System,4 which layabove the Silurian. This wascharacterized, as its name suggests, byparticular rocks in Devon, southwestEngland.

Murchison’s speed

In naming the Permian System inDecember 1841, Murchison perhaps felthe had to move fast. He gave a clue tohis motivations in his PresidentialAddress to the Geological Society ofLondon a few months later, in February1842. In the published version,5 hewrote:

Whilst on the topic of Russia, I will now state,that if on account of the preparation of thisdiscourse and other official duties I had notbeen greatly occupied, I might before now havepresented to you some of the results of thesecond visit to that country.… We offer … thechief results of our enquiries, and … placethem on record as bearing date from September1841.

Among the results I will now merely alludeto the first announcement of some of them, in a

letter of the above date, addressed to DrFischer de Waldheim in Moscow, in which …were … the classification of certaincupriferous deposits of sand, marl, limestone,&c, under the term of ‘Permian system’.… Ishould not now occupy your time by alluding toit, had not the mention of the word alreadycalled forth from M. A. Erman the remark, thatthese deposits have been long known to otherobservers.

Murchison goes on, after somejustification of the originality of hisobservations and of his new name,Permian, to say that ‘I was, therefore,surprised to read the premature criticismof M. A. Erman’.

Georg Adolf Erman (1806–77) wasa German physicist and explorer whohad travelled widely through Russia,

Siberia and Kamchatka. Murchisonclearly felt slighted by some of hisremarks and sought to ensure priority inprint for his term, Permian, in case someother geologist beat him to it. As in allthings, in science priority is critical, andpriority has to be established by the dateof publication. A letter to a friend, or aspeech at a scientific meeting are notadequate – the new idea or name has toappear in print. Two years previously,Sedgwick and Murchison hadestablished the Devonian System in ashort note in the same PhilosophicalJournal, again in haste, seemingly toensure their priority over other potentialcompetitors.

Conventions were certainly different

then. Scientists such as Murchisonexchanged formal letters describing theiractivities, and it was common for theseletters to be presented to the scientificsocieties of other countries, such as theGeological Society of London. In thisway, the gentlemen scholars who met todiscuss the latest ideas in the meetingrooms of their societies in London,Paris, Berlin or Moscow could keep upto date with new ideas from overseas.

Murchison’s speed of publication,and the desire for priority to beestablished, is clear, though whether hefeared Erman, or some other scholar,cannot now be established. So whilethere is no doubt that Murchison rushedinto print with a half-digested account of

his new Permian System (Fig. 4), he didnot act with unusual haste, judged by themodi operandi of his day. Such informalabstracts or outlines of new ideas werethe normal means of communication,often preceding the full-scale evidence,presented in the form of an illustratedmemoir or monograph, by many months,or even years, if it appeared at all.

4 The title-page of Murchison’sbrief paper in which he establishedthe Permian System.

The Geological Society ofLondon

The Geological Society of London6

existed as a kind of gentleman’s club,but one with a serious purpose. Themeetings were convivial, and were heldevery two weeks during the winterseason, November to June. From July toOctober, the fellows might be out ofLondon, either engaged in field work orsimply visiting their country estates.

The purpose of the society wasgenuinely to promote geologicalknowledge, and the fellows wereexpected to make original observationsand bring accounts of their discoveriesto the meetings. It was not merely atheoretical debating society, however.Indeed, the Geological Society ofLondon, from its earliest days,established a format for their meetingsthat became the accepted normelsewhere. Members were expected togive well-prepared papers and to submitthemselves to incisive discussion andcriticism, both at the meeting andafterwards. Nothing would be publishedby the Society that had not been vettedcarefully: the Society was keen to

uphold its reputation for high standards.This is very much the usual practice inall sciences today, but it was less so inthe 1830s.

The discussion meetings werereported in brief form in theProceedings of the Society, as well asin a number of gentleman’s weeklymagazines, and sometimes in journalsoverseas. When circumstancespermitted, some of the papers wereworked up into full-scale memoirs forpublication in the Transactions of thesociety. But many authors chose not toprepare such full accounts, preferring tomove on to new topics. Others, such asMurchison, however, were perhapsmore serious, and they wanted to

preserve a proper account of their workfor posterity.

A memoir was a major undertaking,requiring the preparation of a full textpresenting the evidence and arguing thecase. A key problem then was the cost,both of printing and of illustrations.Geology is essentially a visual science.You cannot work as a geologist withoutmaps, diagrams, cross sections anddrawings of rocks and fossils. Thegeological author in the 1830s had toengage an artist to produce detaileddrawings, and then an engraver toconvert the drawings into printing plates(lithographs). The Society bore some ofthe expense, which was covered by ahigh subscription rate for the fellows

(three guineas, or £3.15 per year), butthe authors had to contribute as well.Only a wealthy man, like Murchison,could seek to publish memoirs regularly.

Putting the rocks in order

So how had Murchison arrived at hisnew Permian System? Murchison’smotivation in the 1830s and 1840s wasto sort out the complexity of the geologyof the world. A large task, andinconceivably grand for one persontoday. However, in his PresidentialAddress to the Geological Society ofLondon in February 1842, as well aspresenting his new Permian System, he

enlarged a little on advances in geologyduring the previous year, touching onnew discoveries from every continent.He surveyed progress up to 1841 inestablishing an international scheme forthe division of geological time:7

I am encouraged to hope that the word‘Silurian’, which has been warmly sanctionedby the classic authority of Von Buch, which E.de Beaumont and Dufrénoy have engraved upontheir splendid map of France, and which ourfellow-labourers in America have adopted, willnot be obliterated to make way for other nameswhich are not founded upon any newdistinctions, stratigraphical or zoological. Solong, gentlemen, as British geologists areappealed to as the men whose works in the fieldhave established a classification, founded onthe sequence of the strata and the imbedded

contents, so long may we be sure that theirinsular names, humble though they may be,will, like those of our distinguished leader,William Smith, be honoured with a preferenceby foreign geologists.

An arrogant, insular view perhaps, butstartlingly prescient. The geologicaltime-scale which Murchison and othershad been drawing up in Britain did gaininternational acceptance. Everyonetoday uses Murchison’s terms Silurian,Devonian and Permian to refer to rocksof a particular age, containing certainfossils, all round the world, not just inBritain, Germany, France and America,but from Argentina to Alaska, Australiato Azerbaijan.

Murchison clearly had a mature

vision in 1842, but the earlydevelopment of the geological time-scale had been somewhat haphazard.8Miners and quarrymen had long givennames to the rocks they worked forcommercial advantage. Coal came froma sequence of rocks called the CoalMeasures by British miners, or‘carbonifère’ (‘coal-bearing’) by Frenchworkers. We have already encounteredanother of these miners’ terms, theKupferschiefer of Germany, a term thatis still used locally to refer to the lowerpart of the Zechstein succession of theUpper Permian. Terms such as thesewere certainly utilitarian, but they didnot say anything about the history of theEarth, nor could they be applied on an

international scale.

Mr Smith’s practicalstratigraphy

Murchison, in the quotation above,clearly identifies William Smith as ‘ourdistinguished leader’, and he is oftennow called the father of stratigraphy.Murchison spoke with added affection,since William Smith (1769–1839) haddied only three years earlier, and he wasright to honour him in this way. Smithwas a working geologist, who earned hisliving surveying the routes for canalsthroughout England. He walked long

distances, assessing the rocks that wereto be cut through, estimating the costs ofthe workings and identifying potentialproblems for his paymasters. Murchisonhad first met Smith, then aged 57, inYorkshire in 1826, when Smith showedthe young tyro the rock successionsalong the coast.

As he traversed England, Smith hadnoticed certain repetitions in the rocktypes and in their contents. The rocksseemed to occur in recognizablesequences, and certain rocks containedpredictable assemblages of fossils.From these observations he laid some ofthe key building blocks for geology as anorganized science. Smith realized thatfossils could be used to identify different

ages of rocks, and that it would bepossible to apply such schemes overlong distances, perhaps evenworldwide, to produce a standard viewof the order of rocks, and thus the orderof events in geological time. As Smithwas aware, this would be anindispensable tool for the geologicalsurveyor, who could enter a previouslyunexplored area, search for fossils,consult a standard compendium and thengive a date for the rocks in front of him.In Smith’s day, wealthy landownerswere sinking boreholes on their landhaphazardly in the search for coal. Intheir ignorance of the ages of the rocks,most of these boreholes were a completewaste of time since they penetrated

rocks that were either too old or tooyoung.

Unlike Murchison, Smith was not akeen writer; while he had just aboutformulated his model of the stratigraphyof England by 1799, he did not publish ituntil much later. In particular Smith hadsorted out the Jurassic and Cretaceous,and some of the divisions of these majorsystems, based on their rich fossilcontent. For example, in the MiddleJurassic around his home district of Bathand the Cotswolds, different rock unitscould be distinguished based on theirparticular assemblages of brachiopods(‘lamp shells’), bivalves and ammonites.Modern geologists divide the MiddleJurassic into 11 or 12 major zones (Fig.

5), based primarily on ammonites,coiled molluscs that belong to a groupnow extinct, but related to modernsquids, octopuses and the shelledNautilus.

Smith could determine the lower partof the Inferior Oolite by the ammonitenow called Ludwigia murchisonae(named, as it happens, after Murchison),and upper parts by Stephanocerashumphriesianum and Parkinsoniaparkinsoni respectively. The overlyingLower Fuller’s Earth of Dorset andSomerset contains very few ammonites,but is characterized instead by the smalloyster Ostrea acuminata. And so it goeson, bed after bed. Smith could verify thesequence of rock types by observation in

quarries and canal cuttings: everywherethe Lower Fuller’s Earth lay on top ofthe Inferior Oolite, and never the otherway round. And everywhere the upperparts of the Inferior Oolite contained thea mmo ni t e Parkinsonia parkinsoni,among other characteristic fossils.

5 The basis of stratigraphy, withan example from the MiddleJurassic of England, as establishedby William Smith. Two measuredsections are shown, from Chideock(left) and Burton Bradstock (right)in Dorset. Lines linking the twosections indicate correlations. Themain zones and some key fossilsare shown.

William Smith had laid thegroundwork, and during the heady daysof the 1830s, his dream of a universal,worldwide scheme for the division ofgeological time was taking shape.Already by 1830, Murchison, amongothers, had extended William Smith’sstratigraphic work from southern

England to other regions – to Scotland,to France and Germany, and even furtherafield. Geologists hardly dared hope thatthe system might work worldwide, but itseemed to. With reports from theAmericas and further regions of Europe,Smith’s Jurassic succession seemed tobe repeated. The same was true of theother units. This opened up vastpossibilities for the furtherance ofgeology as a science, and it pointed theway for scholars to adopt aninternational outlook.

The New Red Sandstone

Murchison spent most of the 1830s

sorting out the Silurian and Devonianperiods. In doing so, he and other fieldgeologists had resolved one of two greatred sandstone units. These seemed tosandwich the Carboniferous, thesuccession of thick marine limestonesand coal-bearing continental sedimentsthat had long been established. Belowthe Carboniferous across much ofBritain lay the Old Red Sandstone, andabove lay the New Red, both of themcomposed of thick sandstones thatappeared to have been deposited byrivers and lakes in tropical conditions.Murchison argued strongly that the OldRed Sandstone was equivalent in age tothe marine Devonian of the county ofDevon, and history has proved him right.

But what of the New Red Sandstone?The Carboniferous, lying between theOld and New Red Sandstones, with itseconomic coal deposits, was easy torecognize throughout Europe, andMurchison was not interested in goingover old ground. The New RedSandstone in turn was overlain by theJurassic, which William Smith and othergeologists in England, France andGermany, had long ago worked out.Again, this was old news, andMurchison was not interested. The NewRed Sandstone was for Murchison theonly substantial gap in the globalstratigraphical scale that he and otherswere rapidly completing. To plug thatgap would be to complete the great

international programme of a singletime-scale.

The upper parts of the New RedSandstone had been termed the Triassic(‘three-part’) in Germany, by Friedrichvon Alberti (1795–1878), Inspector ofSaltworks at Friedrichshall, southwestGermany. Alberti chose the namebecause the German Triassic wascharacterized by a three-part sequence: abasal unit of red and yellow sandstonesand conglomerates called theBuntsandstein (‘coloured sandstone’); amiddle marine unit of limestones andmudstones called the Muschelkalk(‘shelly-limestone’); and an upper unitof red and purple mudstones andsandstones called the Keuper (an older

quarryman’s term for the vari-colouredmudstones).

The German Triassic overlies asequence of marine limestones,mudstones and evaporites (salts) calledthe Zechstein (an older miners’ term),which in turn succeeds a succession ofred and yellow sandstones termed theRotliegendes (‘red beds’). TheRotliegendes and Zechstein weretentatively identified in England, but itwas hard to mark their boundariesprecisely. The tripartite GermanTriassic could not be recognized sincethe Muschelkalk appeared to be missing.In the end, the British geologists wereforced to continue to use the ratherfeeble, informal term New Red

Sandstone.Despite this lack of clarity, another

British geologist had already recognizedthat something major was happening inthe middle of the New Red Sandstone.Between the undefined post-Carboniferous sandstones andlimestones of the lower part and theGermanic Triassic of the upper, a hugechange had taken place in the typicalanimals found in the sea. All the oldtypes had gone and new ones hadappeared. This geologist was JohnPhillips, as energetic as Murchison, butfrom a very different background.

John Phillips and the valueof fossils

John Phillips (1800–74) was of a lowersocial class than Murchison orSedgwick.9 But, by his energy and goodsense, he was able to achieve highregard in his lifetime among fellowgeologists. Phillips was in fact thenephew of William Smith, and helearned a great deal of practical geologyfrom his uncle as they walked aroundSomerset and Wiltshire. As a youngman, and with no formal training,Phillips was offered a museum post inYork in 1824, where he was invited torearrange and display the fossil

collections. It was here that Murchisonmet Phillips in 1826, while he wasengaged in this task. Phillips carried outsimilar curatorial work elsewhere, andhe was secretary of the BritishAssociation for the Advancement ofScience from its inception in 1831. Herose from these somewhat menialpositions to become Professor ofGeology successively at King’s College,London in 1834, Trinity College, Dublinin 1844, and finally at Oxford in 1856.

Phillips studied mainlypalaeontology and geology, but he alsomade contributions to meteorology andastronomy. He wrote a number of bookson the geology of Yorkshire, the Oxfordregion and southwest England, as well

as a series of highly successful textbooksaimed at keen amateurs, of which therewere thousands, and school anduniversity students. Phillips was arecognized palaeontological authority,and he argued repeatedly for theimportance of basing stratigraphy onfossils. In other words, he insisted thatunits of geological time should befounded on their fossil content, andhence made internationally applicable.

There was no point, Phillips argued,in establishing a geological system for amerely local manifestation of rocks,perhaps only to be seen in Devon or innorth Wales. Geological systems shouldbe defined, top and bottom, by uniquefossils that could be recognized readily

wherever geologists went in the world.In this view, Phillips was upholding hisuncle’s classic work. Phillips influencedSedgwick and Murchison deeply, andthese essentially field geologists werefully convinced of the need to submit thefossils they found to expertpalaeontologists, and to rely on theirjudgments in correlating rocks fromplace to place.

The end-Permian massextinction recognized?

Phillips produced a famous diagram(Fig. 6) which is often taken as the first

recognition of the great end-Permian andend-Cretaceous mass extinctions. Heshows the diversity of life by the curveswelling out to the right. But mostimportant to our eyes are the two dropsin diversity. Surely here Phillips meantto show the two greatest cataclysms inthe history of life at the end of thePermian and Cretaceous periods?

6 A diagram showing JohnPhillips’ conception of thePalaeozoic, Mesozoic andCenozoic, as shown in his textbookof 1860. The wavy line broadlyrepresents the diversity of life, andthe boundaries between thePalaeozoic and the Mesozoic andthe Mesozoic and the Cenozoic areclearly indicated as times ofreduction in diversity and turnoverof faunas.

Not in the modern sense. Phillipsintended the diagram – which appearedin one of his books in 1860 but surelyrepresented his opinion in 1840 also – toshow how geological time could bedivided into three major units, or eras.The eras were more all-encompassing

than the geological periods thatMurchison and Sedgwick wereestablishing: the Silurian, Devonian andPermian were merely parts of thePalaeozoic Era, for example. Phillips’point was that the three eras could berecognized by quite differentassemblages of fossils, and this wouldapply worldwide. The faunas and florasrepresented by ‘Palaeozoic life’ and‘Mesozoic life’ were quite different, andlinked by only the most tenuoustransitional forms, but geologists andpalaeontologists were not ready then forevolution and mass extinction as weunderstand them today.

Sedgwick had in fact named thePalaeozoic Era in 1838, and he intended

this new unit to include all the pre-Carboniferous rocks that he andMurchison were then tussling over – inother words their Cambrian andSilurian. Phillips immediately graspedSedgwick’s new concept, and extendedit, in articles he wrote in the PennyCyclopedia in 1838 and 1840. Phillipswas not a rich man, and he was happy tobe paid for such hack writing, while thereaders of the Penny Cyclopediaprobably did not appreciate theoriginality of what they were reading.

For Phillips argued that the wholegeological column could be divided intothree great series, or eras, the first beingSedgwick’s Palaeozoic. The second wasthe Mesozoic and the third the

Kainozoic, both of which Phillips namedin one of his Penny Cyclopedia articles.These three terms are pure Greek,meaning respectively ‘ancient life’,‘middle life’ and ‘recent life’, inreference to the relative antiquity of theirfossils. The Kainozoic came to be morecommonly spelled Cainozoic, sometimeseven Caenozoic, but it is now generallyrendered as Cenozoic. Each of the threedivisions was then characterized byutterly different suites of fossils.

Phillips argued that Sedgwick’sPalaeozoic could not be restrictedsimply to the Cambrian and Silurian, butthat it had to be extended to theCarboniferous, since the fossils of thenew Devonian system supplied a faunal

link between the Silurian and theCarboniferous. Phillips went further,proposing that the Palaeozoic shouldextend into the New Red Sandstone.

This idea was revolutionary, and itplaced a strong focus on the New RedSandstone. If, as Phillips said, thedivision between the Palaeozoic andMesozoic eras lay somewhere withinthis succession of red sandstones, whereshould it be placed precisely?

Murchison overseas

Despite Phillips’ contributions, the NewRed Sandstone still stood in everyone’s

way. Without a clear definition of thesequence of rocks between theCarboniferous and the Triassic, anyinternationally applicable system of timedivision would fail. Murchison realizedthat he could not sort out the New RedSandstone in England, and he hadalready seen the Rotliegendes andZechstein and the classic Triassic ofGermany and central Europe. He wouldhave to venture overseas again to findsomewhere that displayed a completesequence of rocks lying between theCarboniferous and the German-styleTriassic.

Such an international endeavoursuited Murchison well. He knew that thescientific leaders of the new Victorian

age (Queen Victoria had acceded to thethrone in 1837) would have to be worldtravellers, people with friends in all themajor scientific capitals and who couldspeak several languages. Murchison wassuch a man. He spoke French fluently,happily giving talks to the scientificsocieties in Paris and publishing papersin French from time to time. He alsoknew enough German to get by. Despitehis long visits to Russia, however, hedid not speak Russian. But then theRussian savants did not either: theymuch preferred to speak and write inFrench, or, failing that, German orEnglish – anything other than thebarbarous tongue of the serfs andpeasants.

Murchison visits Russia

Murchison’s first visit to Russia was notin fact to solve the New Red Sandstoneproblem. He was still reinforcing hisSilurian and Devonian systems, andhoped to find new evidence to resolvethe matter once and for all. Sedgwickand Murchison had geologized togetherin northern Germany and Belgium in thesummer of 1839, and they had alsovisited Berlin and eastern parts ofGermany. In the written account of theirexplorations, they repeatedly used theirobservations to confirm the validity oftheir recently established Cambrian,Silurian and Devonian systems.10 Theirrapid traverses over vast tracts of

country tempted Murchison to headfurther east. He decided on an excursionacross northern Russia in 1840, theplanning for which occupied all his timethrough the preceding winter.

Murchison gave several briefaccounts of this first trip to Russia. Hespoke at the British Association meetingin Glasgow in September 1840, and thengave a joint presentation to theGeological Society of London with hisFrench colleague, Edouard de Verneuil,in March 1841.11 In their papers,Murchison and Verneuil described theextent of the information that had beenavailable to them before their visitconcerning what they might expect tofind in northern Russia. In fact, there

was essentially a single paper,12 a shortmemoir in the Transactions of theGeological Society of London,published in 1822, with the remarkabletitle ‘An outline of the geology ofRussia’ and a mere 39 pages long. Theauthor was W. T. H. Fox-Strangways,the then British Ambassador in StPetersburg.

Strangways had written a highlycompetent account, including a map, butit was based on only limitedobservations that he had been able tomake in the course of his official dutiesaround St Petersburg, and he was not aprofessional geologist. Murchison andVerneuil had also seen the accounts ofRussian fossils by J. G. F. Fischer von

Waldheim, C. H. Pander and E. I. vonEichwald, but these memoirs said verylittle about the geology on the ground.On the whole, then, they had very littleto go on when they entered Russia.

Murchison and Verneuil began theirtraverse at St Petersburg, at the head ofthe Baltic, where they confirmed that therocks were Silurian in age, overlaid byDevonian. They then went on to describethe Carboniferous, New Red andJurassic rocks that they identified intheir northwards sweep to Archangel,and on their return southwards, acrossthe Volga, and round through NizhniiNovgorod and back to Moscow, ajourney totalling some 6000 kilometres.

In a note added to their paper, dated

26 March 1841, Murchison gives a briefreport of some new findings in thesouthern part of European Russia madeby his Russian colleagues, Baron A. vonMeyendorf and Count A. Keyserling. Healso indicates that he and Verneuilplanned to visit these tracts, as well as,significantly, the Ural Mountains and theregion around Orenburg. When this hadbeen accomplished, Murchison notes,‘all the chief facts will have beenobtained for the construction of a generalgeological map of Russia in Europe’.Such splendid confidence! And he meantit. In 1840, with almost no previouswork to build on, it was necessary toconstruct maps of huge territories on thebasis of such rapid traverses. The detail

could be added later.

Murchison’s second visit toRussia

Murchison’s second visit to Russia hadtwo main purposes. He wanted tocharacterize his new Permian Period,and determine its exact relations to theCarboniferous below and Triassicabove. He also wanted to cement hissocial relations with the Tsar and all hisprinces, and secure funding from themfor his trip and for an ambitiouspublication he had in mind.

The published materials describing

Murchison’s second visit to Russia wererather more extensive than for the first.13

As we have seen, he announced his newsystem, the Permian, in the notepublished in the PhilosophicalMagazine and elsewhere in late 1841.The first public presentation was givenin his Presidential Address to theGeological Society of London on 18February 1842, in which Murchisondefended his new name ‘Permian’ asthoroughly original – and refutedErman’s suggestion otherwise.

The long-awaited full account ofMurchison’s and Verneuil’s second tripacross Russia was read at twoconsecutive meetings of the GeologicalSociety, on 6 and 20 April 1842. In it,

they reviewed the rock sequence inorder, as before, from Silurian toCretaceous, with some comments on theoverlying Pleistocene. The mode ofargument was also as before: Murchisonrecounted the rock types they had seen,and he used Verneuil’s identifications offossils to fix the ages by comparisonwith the sequences in England, Franceand Germany.

Murchison added some furtherexplanation of the Permian System: heequated the Russian Permian with theZechstein of Germany (alreadyrecognized as part of the New RedSandstone) and with the MagnesianLimestone of England. These last twohad long been regarded as equivalents.

In Russia, Murchison noted the huge areacovered by his new Permian, from Permitself to Orenburg in the south, coveringmany hundreds of kilometres of thewestern side of the southern Urals. Hehad followed the sequences in the Uralsup through undoubted Carboniferousrocks, as indicated by the presence ofcoal, but more importantly by thepresence of specific fossils (marineshells and terrestrial plants) shared withthe Carboniferous of western Europe.Above the Carboniferous he had seenPermian limestones, gypsum and saltbeds, red and green sandstones, shales,conglomerates and copper-richsandstones (the miner’s so-called‘Kupferschiefer’).

Bottom and top of thePermian: molluscs and

reptiles

Murchison knew he had to do two thingsto identify the lower New RedSandstone as an acceptable internationalstratigraphical system, his Permian.First, he had to define the base, andsecond he had to find characteristicfossils that placed the age of the rockunits definitively between theCarboniferous and the Triassic. Instratigraphy, defining the base of asystem is crucial. The top doesn’t mattersince it is defined by the overlying unit.In this case, the work had already been

done, since the base of the Triassic hadbeen fixed in Germany by Alberti whenhe named that system in 1834, as we sawabove.

Murchison had studied the gypsum-and salt-bearing lower New Red unitsaround Perm, and Keyserling hadprovided him with the relevantinformation on the fossil-bearing rocksaround Arti. These were marine rocks,full of shells and other, identifiablefossils which allowed Verneuil to makedirect comparisons with well-studied,definitive Carboniferous and Triassicrocks of western Europe. So Murchisonhad the Carboniferous/New Redtransition documented south of Perm inmarine rocks with fossils. He had seen

the brachiopods and plants himself, andhe now had the clincher, bones of fossilreptiles that Richard Owen haddetermined were truly intermediate incharacter between Carboniferous andTriassic forms. The shells, the plantsand the bones all seemed to confirm hisconclusion – the Russian Permian laybetween the Carboniferous and theTriassic. He had plugged thestratigraphic gap.

This conclusion is borne out by thefamous letter Murchison wrote toFischer von Waldheim when he reachedMoscow in October. In it he justified hisnew Permian system on the basis of thefossils, some of them Carboniferous inappearance (brachiopods), and others

intermediate between Carboniferous andTriassic (plants, fishes, reptiles):14

Of the fossils of this system, someundescribed species of Producti [brachiopods]might seem to connect the Permian with thecarboniferous aera; and other shells, togetherwith fishes and Saurians, link it on moreclosely to the period of the Zechstein, whilstits peculiar plants appear to constitute a Floraof a type intermediate between the epochs ofthe new red sandstone or ‘trias’ and the coal-measures.

These arguments were repeated in laterpapers, but Murchison had clearlyformulated his interpretation while hewas in the field near Orenburg. On 20November 1841, less than three weeksafter his return to England, and just over

a week before the name ‘Permian’ waspublished in the PhilosophicalMagazine, Murchison wrote to RichardOwen:15

I have been coming to you for some daysbecause I wish you to see the bones of a saurianor two from the lowest deposits in which suchanimals are found in Russia, and which I believeto be on the level of our MagnesianLimestone…. I am personally much interestedin the determination of these saurians as theyes or no to a certain query which may guideme very much in my Imperial classification.

The yes or no question was: are thesereptiles intermediate between those ofthe Carboniferous and the Triassic?Owen’s answer, as we’ve seen, wasyes, and Murchison was vindicated, as

he stressed in his President’s Address tothe Geological Society of London inFebruary, 1842.

The Geology of Russia(1845)

After his first trip in 1840, Murchisonplanned to present the results of hisresearches in Russia simply as memoirsin the Transactions of the GeologicalSociety of London. However, as he sawthe great support he was receiving inRussia, from the Tsar downwards, heconceived the idea at the beginning ofhis second trip of publishing a more

substantial work, a large book thatwould stand beside The Silurian Systemas a worthy successor. Count I. F.Kankrin, the Russian Minister ofFinance, had hinted both before and afterthe 1841 trip that the Russiangovernment would underwrite theproduction of a map and report. Theydid not realize, perhaps, that theirconcept and Murchison’s differedconsiderably in scope and magnitude.

Early in 1842, Murchison plannedhis great Russian book.16 It wouldconsist of 600 pages, divided into threesections, with plates of fossils and alarge map. The first section was to beabout the geology of Russia in Europe,the second on the Ural Mountains and the

third on the fossils. Murchison wrotesolidly through 1842, and delivered thefirst chapters for typesetting inNovember of that year. He continuedsending chapters for typesetting during1843, but the project was clearly gettingout of control. He did further fieldwork:in Poland, Germany and Czechoslovakiain 1843, Denmark, Norway and Swedenin 1844, and these countries plusGermany in 1845. This new work forcedhim to reconsider his earlier chapters,and in the end he had to cancel the first40 pages and replace them with nearer80 pages of new text.

The fossils were another problem.Murchison was not a palaeontologist,and most of the fossil descriptions were

to be done by Verneuil and others(although Murchison himself diddescribe the trilobites). He soon askedVerneuil to arrange for the fossils to bedescribed in a separate volume whichwould be published, in French, in Paris,at the same time as his geologicalaccount. Production of two volumes, intwo languages, by two authors, meantthat there were all sorts of confusionsand mistakes, but the work had to bepushed forward.

So in the end the planned 600-pagevolume had more than doubled in size.There were two volumes, totalling over1300 pages, 60 plates and two largemaps. Everything was finished in a rush.Murchison’s volume 1 bears the date

‘April 1845’, but printing of revisedpages and illustrations continued afterthat. Verneuil’s volume 2 beganproduction in 1844. In a great rushduring June and July 1845, six copies ofthe complete work were put together,and Murchison carried them to StPetersburg on 29 July 1845. Hepresented the six sets to Tsar Nicholas Ion 12 August 1845 who studied themwith evident approval and intelligentappreciation.

The end-Permian massextinction ignored

Thus the years 1840 and 1841 marked acritical time for the understanding of theend-Permian mass extinction. Phillipshad established the exact extent of thegeological series, the Palaeozoic,Mesozoic and Cenozoic, and he hadmade a firm link between these majortime divisions and extinction. Murchisonhad named the Permian System, and hehad shown the importance of Russia inunderstanding that span of time.However, neither Phillips, norMurchison, nor any other geologist of theday, pointed to the huge mass extinctionthat had occurred at the end of Permiantime.

This omission can be explained intwo ways, one rather simple, the other

perhaps more fundamental. It couldsimply be that the geologists andpalaeontologists in 1840 did not haveenough evidence to pin the extinctionevent down. After all, the Permian hadonly just been named, and Phillips hadonly just begun to document large-scalechanges in the fossil record through time.More fundamental, though, had been ahuge change in mindset that had comeabout during the 1830s. Whereas in 1830it had been legitimate to talk aboutcatastrophes in the geological past, by1840 catastrophism had become a dirtyword. This astonishing shift ingeological philosophy had beenengineered by one man, and his effortsset the tone of geological enquiry for

150 years.

3

THE DEATH OFCATASTROPHISM

I learned my geology at the University ofAberdeen in the 1970s. Our first-yearlectures were delivered by a magnificentScottish professor of the old school,who covered everything, from theformation of the Earth to the ice ages,from fossils to mineralogy. These were

the days before colour slides wereubiquitous, and he laid out largedrawings made on white boards aroundthe room before his lectures began.Assiduous students arrived early to copythe complex diagrams into theirnotebooks.

I remember his introductory series oflectures, which covered the history ofgeology. He told us about Smith,Murchison, Sedgwick and Lyell, whohad laid the foundations of the subject.These men had overcome every obstacleto reveal to us the truth of the history ofthe Earth. They allowed no mysticism,no wild notions, none of the strangeramblings that dominated the earlynineteenth-century scene. These fine,

upstanding geologists, fathers of thediscipline, British through and through,had blazed a trail through theobscurantism and fog emanating from thegeological societies of Germany andFrance. Indeed, not only were theyBritish, but two of them, Murchison andLyell, were Scots. Best of all, the leaderof this group of reformers, CharlesLyell, had been born only a few milessouth of Aberdeen. Scottish commonsense had prevailed over foreignimagination.

Such was the story our professortold us – as it had been told to countlessgeologists over the years. I happilyaccepted it. On reflection, though, thesesimple stories are usually more

complex. Can it really ever be the casethat one group of philosophers orscientists is completely right, and theother completely deluded? Clearly not.

What had happened, it now seems,was that Lyell and his supportersrewrote the history of geology. In thebest political tradition, geologistsbecame divided sharply into two camps,and labels were applied. The good guyswere the uniformitarians, and thebaddies were the catastrophists.Murchison, Sedgwick, Phillips, andindeed most geologists in Britain, wereactually catastrophists, but Lyell wasable to wield such influence that they allrather cravenly accepted his labels andhis strictures. Anything but be classified

as a catastrophist!And yet the catastrophists were right

about extinctions. It is crucial tounderstand this dramatic switch ingeological reasoning which happenedaround 1832. Virtually overnight, itseemed, Lyell engineered a newunderstanding in geology and, in doingso, he made it very hard for any rationalperson even to think about massextinction. He was the victor, and hisvision dominated geology for over 150years.

Murchison and theories ofthe Earth

Roderick Murchison was a practicalfield geologist and he did not venture agreat deal into theorizing about broadissues, about the history of the Earth, theprogress of life or the meaning of time.But he lived through two major debates,both masterminded by Lyell and both ofwhich made him rather uncomfortable.First, did time run in a single directionor not, and second, was the history of theEarth dominated by catastrophes or not?

Charles Lyell (1797–1878), suaveand impressive, was five yearsMurchison’s junior. He was the authorof an influential textbook, published inthree volumes between 1830 and 1833,entitled Principles of Geology,1 whichachieved iconic status and is still

regarded by most people as one of thefounding volumes of geology. And yetthis epochal volume contains someunbelievable nonsense among its goodmaterial. For instance, Lyell wrote thatif temperatures were to increaseworldwide, ferns and other primitiveplants might clothe the lands, and

Then might those genera of animals return, ofwhich the memorials are preserved in theancient rocks of our continents. The hugeiguanodon might reappear in the woods, and theichthyosaur in the sea, while the pterodactylemight flit again through umbrageous groves oftree-ferns.

In his Principles, Lyell makes it clearthat he sees groups of plants and animalsas eternal. They come and go simply in

response to physical conditions on theEarth. This seems incredible, but Lyellwas not to be disputed in the 1830s and1840s. Extraordinary as his commentsmight seem, most geologists acceptedthem. Perhaps in private they doubtedhis views, but they did not speak out inpublic.

The other debate that influencedMurchison was that betweencatastrophism and uniformitarianism.Lyell had pinned his colours to theuniformitarian mast in the subtitle of thePrinciples: ‘being an attempt to explainthe former changes of the Earth’ssurface, by reference to causes now inoperation’. He argued strongly, andrepeatedly, that every geological

phenomenon could, and should, beexplained by modern processes. This isthe principle of uniformitarianism, ‘thepresent is the key to the past’, and itseems eminently sensible. Surely it isbetter for a geologist to refer to the rulesof physics, and to the observations ofmodern geographers and biologists, thanto imagine unknown processes?

French and other continentalgeologists, on the other hand, were at thesame time postulating explosions,meteorite impacts, sudden extinctionsand miraculous events in the geologicalpast. This is catastrophism, and clearlyit must be wild, non-scientificspeculation. Uniformitarianism was thestraightforward, common-sense, British

viewpoint. Lyell had the rightcredentials to be considered sensible,twice over, as both a Scot and abarrister. To deny him, and to denyuniformitarianism, would be to labeloneself a wild theorist.

Murchison could not risk such aderogatory label, although he had been acatastrophist himself in the 1820s. He,and most other geologists at the time,now strove to avoid the ‘catastrophist’label by denying all reference to suddenevents or major calamities in the past.They also accepted both Lyell’s doubtsabout the extinction of major groups andthe possibility of continuous recycling ofgroups on and off the scene, althoughmany found this unpalatable. In

following Lyell, Murchison ultimatelyhad to deny the possibility of massextinctions. He also had to reject theviews of the highly respected GeorgesCuvier, the greatest catastrophist of theday.

Cuvier and catastrophism

There was nothing new in catastrophismfor Cuvier. Such notions had beenaround since the seventeenth century,when savants speculated about the originof the Earth, and the nature of planets,comets and meteorites.2 Cuvier waslater unfairly bracketed with those early

speculators. Yet he was always thoroughin marshalling his evidence before hedrew any conclusions, and indeed hehimself at first mocked the geologists ofhis day as wild speculators who basedtheir ideas on minimal evidence.

Then in around 1800, Cuvier beganto study the fossil mammals found in theParis basin. Beautiful completeskeletons of small horse-like and dog-like animals had been excavated fromthe Eocene and Miocene limestones, andCuvier could see that the skeletons wereclearly those of extinct forms. Hedescribed the circumstances of theiroccurrence in a report in 1810:3

The unknown bones are almost always covered

by beds full of seashells. It is thus some marineinundation that annihilated the species; but theinfluence of this revolution, by its very nature,was not perhaps exercised on all marineanimals.

Here he uses the word ‘revolution’ tomean some major physical calamity. Hecomes back to this question a few pageslater:

The primitive diminution of the waters, theirrepeated returns, the variations of the materialsthey deposited and which now form our strata;those of the organisms whose remains fill apart of these strata; the first origin of thesesame organisms: how are such problems to beresolved with the forces that we know now innature? Our volcanic eruptions, our erosions,our currents, are pretty feeble agents for suchgrand effects.

Cuvier did not have strong evidence atthis point, nor was he precise about thenature of the ‘revolutions’, but he wasclearly positing processes, and scales ofprocesses, operating in the past that donot operate today.

Après moi le déluge

Cuvier developed the theme ofrevolutions in much more detail in hismore famous ‘Discours préliminaire’ tohis great book Recherches sur lesossemens fossiles, published in severaleditions from 1812 onwards.4 Thepreliminary discourse was a fuller

manifesto than the 1810 report, outliningfor the general public Cuvier’s views ongeology and evolution, and it was hugelyinfluential, being translated and reissuedmany times, even long after Cuvier’sdeath. In the fifth section of thepreliminary discourse, Cuvier wrote:

Thus the great catastrophes that producedrevolutions in the basin of the seas werepreceded, accompanied, and followed bychanges in the nature of the liquid and in thematerials that it held in solution.… During suchchanges in the general liquid, it was verydifficult for the same animals to continue tolive in it. And they did not do so. Their speciesand even their genera change with the beds.

Cuvier goes on to discuss the physicalprocesses that act on the Earth today,

and he makes a strong plea forconsideration of more catastrophicprocesses than these in the past. Thecatastrophes seem to have affected landanimals, and especially the mammals,much more than the lowly sea creatures:

the nature of the revolutions that have alteredthe surface of the globe must have had a morethorough effect on terrestrial quadrupeds thanon marine animals. Since these revolutionslargely consisted of displacements of theseabed, and since the waters must havedestroyed all the quadrupeds that they reached,… they could at least have annihilated thespecies peculiar to those continents, withouthaving the same influence on marine animals.

I was intrigued to discover this passage,since it resonated strangely with some of

Murchison’s remarks in his PresidentialAddress to the Geological Society ofLondon in 1842.5

Cuvier further clarified his views onthe rapidity of the repeated inundationsand retreats of the seas in later editionsof the preliminary discourse:6

the majority of the cataclysms that producedthem were sudden. This is particularly easy todemonstrate for the last one.… It … left innorthern countries the bodies of greatquadrupeds, encased in ice and preserved withtheir skin, hair and flesh down to our owntimes. If they had not been frozen as soon askilled, putrefaction would have decomposedthe carcasses.… This development was sudden,not gradual, and what is so clearlydemonstrable for the last catastrophe is notless true of those which preceded it.… Life in

those times was often disturbed by thesefrightful events. Numberless living things werevictims of such catastrophes: some, inhabitantsof the dry land, were engulfed in deluges;others, living in the heart of the sea, were leftstranded when the ocean floor was suddenlyraised up again; and whole races weredestroyed forever, leaving only a few relicswhich the naturalist can scarcely recognize.

Much of this is reasonable enough, and ithardly seems to justify Lyell’s completerejection of Cuvier’s geological ideas aswildly dangerous.

It was clear to Cuvier, as to anygeologist today, that the different layersof sediment are often separated byabrupt discontinuities. He had seenmarine limestones in the Paris basin, full

of seashells and fishes, which werecapped by terrestrial mudstones andsandstones containing skeletons of landanimals. Cuvier’s only mistake wasperhaps to try to visualize the rocksequences accumulating more rapidlythan would be assumed by geologiststoday. This meant that he had to imaginethat all sea-level changes, all shifts inthe positions of coastlines, happenedvirtually instantaneously. In reality, it ismore likely that such floodings andretreats of the sea took thousands or tensof thousands of years, as they have donemore recently, in historical time.

Cuvier did not in fact claim that theEarth had been subject to impacts, vastdisturbances or other unusual events. But

he certainly claimed a role for events ona larger scale than are known today. Hissupporters, in France and in Britain,argued that the Earth’s crust had beensubjected to forces in the past that nolonger obtained. Some noted the intensefolding and the evidence of melting ofrocks in the ancient mountain chains ofFrance and Scotland, and in the Alps.Others assumed that the Earth had beenhotter in the past, and more subject toviolent volcanic eruptions. The last ofthe many extinctions to have affected theglobe, indicated by the frozen mammothsand other large mammals that Cuvier hadmentioned, was attributed to Noah’sflood in the Bible. So the catastrophistdoctrine could be tied to biblical

orthodoxy, and for a while, in the 1820sand 1830s, diluvialists such as WilliamBuckland, supporters of the Flood as thelast of many catastrophes to affect theEarth, were in the ascendant.

Cuvier had influenced a generationof geologists, in France and Britain, andMurchison found it hard to adjust toabandoning this stance. But Lyell’scritique in the 1830s was firm andpersistent. He seized on Cuvier’sDiscours as his bête noire, and he didnot give up. Cuvier had preachedrevolutions, evolutions and catastrophes.Lyell decided this was dangerous stuff,and he attacked it fiercely. Geologybooks today celebrate Lyell as the heroand Cuvier as the vanquished. And yet

much of what Cuvier said was right andmuch of what Lyell said was wrong.

Lyell and Murchison gotravelling

Charles Lyell (Fig. 7) was born atKinnordy, Angus, to a landed family. Hewent to Oxford to study law, butpractised as a lawyer unenthusiasticallyfor only two years, before taking up hisgreat love, geology, full time. LikeMurchison, he had moved to London,and worked professionally as ageologist, but supported himself partlywith his private wealth. This proved

insufficient, however, and he relied alsoon income from his numerous textbookson geology.

While at Oxford, he had attendedgeology lectures by the great, buteccentric, Professor William Buckland(Chapter 1 – ‘Giant Saurians’).Buckland was one of Cuvier’s leadingsupporters in England, and a strongproponent of the diluvialist viewpoint.Lyell was deeply influenced, and heswitched to geology in the mid-1820s,about the same time as Murchison alsomoved to geology from his life ofmilitary manoeuvres and fox hunting.

Lyell visited France in 1823 andexamined the rocks of the Paris basin,the very successions that Cuvier used as

evidence for repeated revolutions,advances and retreats of the sea. Heheard some dissenting views, but on thewhole he was prepared to acceptCuvier’s interpretations at that time. In1828, Lyell and Murchison wenttogether on a field trip to the MassifCentral region of France, where theysaw some rock sequences thatapparently made Lyell begin to doubtCuvier’s revolutions. One succession ofmarls, over 220 metres (700 feet) thick,was composed of repeated thin layers,about 12 per centimetre (30 per inch) – atotal of a quarter of a million individuallayers. Lyell recognized the fossils inthese marls from his studies of modernlakes in Scotland and he was able to

conclude that the great thickness ofancient rocks had accumulated in a hugelake over a very long span of time. Butsurely each thin layer could notrepresent a revolution?

7 Charles Lyell in 1836, from aportrait by J. M. Wright.

Lyell and Murchison continued their

wanderings to southern Italy, and theysaw active volcanoes around Naples andin Sicily. This impressed Lyell with thecurrent forces of nature. He could workout the power of the volcanoes to raisethe land up and to deposit vast amountsof lava. In places, he saw beds of rockcontaining marine shells that had beenlifted high above the current sea level,but within historical time. The key tomany of these observations was the factthat there were ancient Romanarchaeological remains among thevolcanoes and uplifted marine beds,proving that the upheavals andcataclysms had happened within the last1500 to 2000 years. Lyell determinedfrom this that modern agencies were

adequate to explain all the phenomenaobserved by geologists, and that therewas no need to call upon large-scaleupheavals as Cuvier had done. When hereturned to England in 1829, Lyellalready had the plan for his book, hisPrinciples of Geology, the first volumeof which appeared in 1830.

Lyell’s advocacy

The strategy of the Principles waseminently laudable, but Lyell introducedconfusion where it did not exist. Lyell’smain claim, throughout the three volumesof his book, was that every phenomenonin geology can be, and should be,

interpreted in terms of processes that canbe observed today – the principleusually termed ‘uniformitarianism’,though this was not his word. He puttogether all the evidence he had seen inhis geologizing rambles round Britain,France and Italy, with reference to othergeological work up to that time.

But Lyell’s Principles is not ageology textbook, as is often thought: itis a legal argument, or a piece ofadvocacy. Indeed, critics at the timerealized what was going on. AdamSedgwick, in his Presidential Address tothe Geological Society of London in1831,7 said that Lyell uses the ‘languageof an advocate’. Another critic, W. H.Fitton said8

The book at first appeared to us to be theproduction of an advocate, deeply impressedwith the dignity and truth of his cause. The tonewas rather that of eloquent pleading than ofstrict philosophical enquiry.

Lyell, then, used his considerableskills as a writer and as a legal debaterto present a powerful argument for anew way of doing geology, and the keywas uniformitarianism. But he mixed upseveral different meanings of uniformityin his book.

The meanings of uniformity

Lyell used the term ‘uniformity’ in at

least four ways in the Principles,9 and indoing so he persuaded the geologicalworld of some very odd conclusions.The first two meanings are certainlysensible and would be readily acceptedby any intelligent field geologist then,and indeed now. But the other twomeanings were something of a distortionof common sense. And, running throughhis argument, Lyell sowed the seeds of afalse dichotomy between the heroes ofuniformitarianism and the dupes ofcatastrophism.

Lyell’s four usages of the word‘uniformity’ are:

• The uniformity of law: the laws ofnature are constant through time; it

would be wrong to suggest that the orbitsof the planets, gravity, mechanics, theinteractions of chemical elements or anyother such fundamental laws of naturehave changed.

• The uniformity of process: the use ofobservations of modern phenomena tointerpret the past, sometimes termedactualism. Geologists should not inventprocesses that cannot be seen today toexplain ancient rock formations.

• The uniformity of rates: processes inthe past must be interpreted to haveoccurred at the same rates as they dotoday, sometimes termed gradualism.So, geologists should not posit massivecatastrophes, floods, volcanic eruptions

or meteorite impacts on a larger scalethan can be observed today.

• The uniformity of state: change onthe Earth has happened in a cyclicalway, and there is no evidence of linear,or directional, change, sometimestermed nonprogressionism. Lyell arguedthat change does not go anywhere, butthat the Earth is in dynamic balance,with climates, volcanic eruptions,deposition of rocks and other phenomenaproceeding at fairly constant ratessomewhere, all the time.

The first two meanings of‘uniformity’ are indeed sensible. Ofcourse a true observational scientist

should use the current laws of nature,and modern processes, in explainingpast phenomena. This is a mixture of theactualistic methodological approach andthe uniformitarian philosophy. Ageologist clearly could not function byspeculating rather than observing, orframing explanations in wild andimaginative theorizing.

But these same messages had alsobeen the solid underpinning ofeverything Cuvier had written aboutgeology from the start. Indeed, Cuvier’sgreat initial distrust of geology andgeologists stemmed precisely from hiscontempt for their wild theoreticalmodels of the Earth, based on minimalevidence. This he repeated time and time

again,10 including in his 1810 and 1812essays. Cuvier also stressed with greatclarity that the geologist must baseeverything on modern processes and themodern laws of nature. So, Lyell andCuvier were in agreement here: bothwere clear thinkers, and both relied ontheir own eyes. But still Lyell somehowsucceeded in casting Cuvier, who wasstill alive in 1830, as the greatspeculator, the enemy of true science.

A claim too far

Lyell’s third and fourth uses of the termuniformity – that rates of processes have

not changed, and that the Earth cyclesendlessly in a nonprogressive way –were statements of belief by Lyell, withno foundation. He simply made these asclaims, and dared anyone to deny him.He presented no evidence, but arguedthem as inevitable extensions of his firsttwo definitions. No change of rate (3)was clearly quite opposite to Cuvier’sviews. Cuvier had argued strongly thatthere had been major revolutions in thepast, and Lyell denied this as wildspeculation. But neither geologist hadevidence, and we now know that Lyellwas wrong. Cuvier, too, was wrong.Each man had pushed his claim too far.

There clearly have been episodes ofvast vulcanism on the Earth, major

movements of the continents, the uplift ofmountain chains, huge meteorite impacts,great extensions of the polar ice caps,and there have been times of quiescence.It just happens that the present is one ofthose phases of modest activity. So Lyellwas completely wrong in his claim forgradualism, or uniformity of rate. Cuviertoo: he pointed to repeated revolutionsmarking every major change of sedimenttype, every postulated advance andretreat of the sea. This is also clearlywrong. But I don’t think any moderngeologist would claim that either Cuvieror Lyell was more or less wrong than theother, and yet Lyell is still regarded asthe voice of reason, and Cuvier as thedeluded, wild-eyed speculator.

Lyell employed the most amazingsleight of hand in trying to convincegeologists of his notion of uniformity ofstate, or his grand cycles of history. Aconsequence of the stately steady-stateEarth was, in his view, that nothing everchanges. Volcanoes erupt, glaciersadvance, sea levels go up or down, buteverything is in equilibrium. He viewedthe Earth as a closed system with certainfixed laws, and that meant that therecould be no progress or direction ofchange. And this he extended to life.Lyell could not in any way accept a timeof origin for life as a whole, nor adecisive point of extinction for anygroup. So, the palaeontologist who triedto say that trilobites or fishes originated

at a particular point in time, and changedin some definite way, and then becameextinct, was denying Lyell’s view.

Quite counter to Lyell’s opinion,palaeontologists in the 1820s and 1830swere coming to realize that there wasapparently some order to the fossils inthe rocks, simple marine creatures first,then fishes, then animals on land, thenreptiles, then mammals. Lyell wouldhave none of such a progressionist view.As quoted earlier, he fully expected tofind fossil mammals in the Silurian, orthat dinosaurs might come back in thefuture. He continued to defend this view,against all odds, well into the 1850s.11

To a certain extent, Lyell was supportedin his anti-progressionism also by

Richard Owen.

Professor Ichthyosaurus

Lyell’s fellow geologists were alwaysuncomfortable with grand cycles of earthhistory, but most of them felt somehowunable to speak their minds. And Lyell’sarguments were so cleverly woven thatit was hard to draw out only thenonprogressionist view of life or of theEarth, and deny them. To query thenonsense was to reject the wholeedifice. Without care, such a stancemight cast the critic into the crowd ofspeculators and creationists who hadbeen so mercilessly dissected and

rejected by Lyell. Murchison, Sedgwickand many other distinguished geologistswere clearly unhappy about some ofLyell’s claims, but they largely keptquiet about it in public.

One trenchant criticism, however, isencapsulated in a famous cartoonentitled ‘Awful Changes’ ( Fig. 8).Professor Ichthyosaurus, wearing apatterned silk coat and a pair of pince-nez, is lecturing his class of marinereptiles, who sit and lie among therocks. In front of the Professor is ahuman skull, about which he says, ‘theskull before us belonged to some of thelower order of animals[,] the teeth arevery insignificant[,] the power of thejaws trifling, and altogether it seems

wonderful how the creature could haveprocured food.’

The cartoon was drawn by Henry Dela Beche (1796–1855), another wealthyyoung man who, like Murchison andLyell, had turned to geology as thecoming science in the 1820s. He wasfamous for his caricatures ofcontemporaries. He drew ProfessorIchthyosaurus in 1831, and distributed itwidely among his friends as alithograph. For a long time this cartoonwas seen as simply a gentle piece of fun,perhaps a spoof on William Buckland,who had indeed written extensively onichthyosaurs.

8 ‘Awful Changes’, a cartoon byHenry De La Beche depictingProfessor Ichthyosaurus lecturinghis class of marine reptiles on theremains of ancient humans.

Ironically, however, ProfessorIchthyosaurus was directed not againstBuckland the catastrophist, but againstLyell the uniformitarian. De la Beche’snotebooks prove the point.12 He wasenvisaging a time in the future, whenichthyosaurs might return to the Earth,very much Lyell’s view, and themselvescomment on the ancient remains ofhumans, in just the same way thatgeologists in the 1820s and 1830s wroteabout the Jurassic ichthyosaurs.Murchison, doubtless, sympathizedentirely with De la Beche’s view.

Murchison’s views

In 1830, when Lyell published hisPrinciples, Murchison was a Cuvieriancatastrophist and a progressionist. Andyet his friend had just issued a polemicagainst both views. What was he to do?The critical response to Lyell was ledby Sedgwick, who attacked Lyell’suniformity of rates, while others attackedLyell’s steady-state non-directionalism.But the opposition was notoverwhelming. Even his critics praisedhis observational examples and hislogical argument: Sedgwick acceptedthat it was wrong to try to link geologyto the Bible through the Great Flood.

Lyell’s book was generally verywell received by the public, despite theprivate mutterings. Charles Darwin

always pointed to the Principles as oneof the key books that influenced his earlywork, even though Lyell’s messageabout the history of life was so entirelyat odds with Darwin’s later views.Somehow, Lyell managed to pull off theamazing trick of pleasing nearlyeverybody, while at the same timeproposing some incredible nonsense.Admittedly, it is easy to be wise now,and Lyell was certainly not the onlygeologist writing highly speculative and,to our eyes, absurd material at that time.But the irony is that his ideas were notrapidly rejected, but on the contrarycontinued to dominate geological thoughtfor decades.

Murchison, perhaps surprisingly,

seems to have stood somewhat in awe ofthe younger Lyell, both in 1830 andprobably throughout his life. Murchisonwas later to become a grandee ofgeology, with a view on every subject,but in 1830 he was yet to prove himself.A young Scottish geologist, J. D. Forbes(1809–68), who later became Principalof St Andrews University but was thenvisiting London for the first time,meeting the great men of his science,wrote some perceptive comments in hisnotebook.13 He stated that Murchison ‘isa diligent observer … but does notappear possessed of much originality’.Lyell, on the other hand, was ‘aremarkable man’, but he was also‘sanguine in his opinions, hasty in his

generalizations’ and always insistent onfinding a theory to explain everyobservation.

Murchison, unlike the other Britishcatastrophists of the day, did not makeany substantial public pronouncementagainst Lyell. And yet, while the secondand third volumes of the Principleswere in press, Murchison already hadenough evidence from his studies of theSilurian rocks and fossils to demolishLyell’s gradualism and hisnonprogressionism. Indeed, in privateMurchison never wavered. In letterswritten in 1851, he commented to acorrespondent that ‘I have repeatedlyshown in other works that operations ofgreat violence, not of Lyell’s quietude,

have been repeated’, and to Lyell hewrote in the same year denying thesignificance of some supposed Silurianvertebrate tracks from North America.14

Lyell’s gradualism of processes, andespecially his nonprogressionism, wereanathema to stratigraphers such asMurchison and Sedgwick – such viewswent exactly counter to William Smith’searlier observation that successions ofrocks were repeated in somewhatpredictable ways, and that each rockformation had its own characteristicpackage of fossils that could berecognized, and used in establishing arelative date. Sedgwick and Murchisonin the 1830s, as we have seen, showedquite clearly that even the most ancient

rocks, the Cambrian, Silurian andDevonian were quite distinct unitscontaining particular fossils. Lyell’snonprogressionism implied that thewhole stratigraphic enterprise, asunderstood then, was wrong-headed.Murchison, however, preferred to mounthis horse and go into the field, and setabout his practical stratigraphic work,leaving the theorizing to others.

The death of massextinction

Lyell’s brilliant disposal ofcatastrophism in the 1830s made it

impossible for anyone to study massextinctions within the bounds of normalscience. At the same time, his viewsundermined the work by Murchison,Phillips, Sedgwick and others, whowere creating the edifice of the standard,global geological time-scale. To thestratigraphers it was clear that fossilschanged through time, and that therewere major gaps and horizons at whichfaunas and floras changed wholesale ingeologically short intervals of time. Butthis had no real meaning in a Lyellianworld of endless cycles of time.

Views about extinction were,however, very different, in 1840.Although, as we saw, the verypossibility of extinction had been stoutly

denied by many scientists until around1800, that bridge had been crossed by1840. But two others had not. It was notclear to Murchison, and to many of hiscontemporaries, that major, co-ordinatedextinctions could occur. And, if suchmass extinctions had happened in thepast, geologists like Murchison andPhillips were unsure about how theycould be linked to known modernextinctions. In light of this, and Lyell’sbaleful and persistent influence, how didgeologists and palaeontologists explainthe growing evidence that catastropheshad happened in the past?

4

THE CONCEPT THATDARED NOT SPEAK ITS

NAME

Today, the end-Permian mass extinctionis generally accepted as having been thebiggest crisis of all time. During thisevent, as we shall see, all forms of life –plants and animals, small and large,terrestrial and marine – were drivenvirtually to total annihilation. And yet, in

1987, in his influential VertebratePalaeontology and Evolution, thestandard text in the field, the Canadianpalaeontologist Bob Carroll wrote:1

The most dramatic extinction in the marineenvironment occurred at the end of thePermian, wiping out 95 percent of thenonvertebrate species and more than half thefamilies. Surprisingly, there was not acorrespondingly large extinction of eitherterrestrial or aquatic vertebrates.

Carroll has since accepted the reality ofthe end-Permian mass extinction amongvertebrates.2 However, his statementreflects a long-standing and respectableposition among geologists andpalaeontologists – the cautious, or

reluctant, acceptance of the reality ofmass extinctions and catastrophes.

A key feature of the debate aboutmass extinctions, from 1840 to 1980, hasbeen the imbalance in perceptions of thetwo sides. The proponents ofcatastrophe and sudden mass extinctionwere consistently regarded as lunatics.To link a mass extinction to cosmic rays,sunspots or meteorite impacts was toclass yourself with the pseudoscientistsand astrologers. The extinction-denierswere the level-headed, carefulscientists. Far better to call for moreevidence, to argue that an extinctionhappened gradually, perhaps over 5 or10 million years, to seek explanation inslow-acting earth-bound processes, such

as sea-level change or climaticdeterioration, than to fly off wildly intothe arms of the soothsayers, doom-mongers and apocalypse-merchants!

Why was mass extinction denied apriori? The explanation almost certainlylies deep within the Cuvier vs. Lyelldebate of the 1830s. Even in the 1960sand 1970s the influence of Lyell’srewriting of history was still there. Andit still is today. But, since 1980, it has atleast become increasingly feasible todiscuss the possibility of mass extinctionand catastrophe without the need to beapologetic. In this chapter, we will tracethe opinions about mass extinctions, bothin general and concerning the end-Permian event, through the years from

1840 to 1980. Here and there, we canidentify some remarkably propheticpublications, in which individuals daredto speak their minds, but at the time theirpapers were ridiculed or, worse,ignored. These were the twilight yearsfor the catastrophists.

Victorian views: problems ofdating

Palaeontologists and geologists afterMurchison barely discussed massextinction at all. They mostly acceptedthat fossil species had become extinct,but those extinctions were seen as

sporadic events, occurring as a normalpart of the ‘life cycle’ of a species.Without any concept of the duration ofgeological time, it was difficult forpalaeontologists in the nineteenth centuryto speak with any conviction about thecoincidence or rapidity of events.

If a nineteenth-centurypalaeontologist had found a layer ofrocks in which 50 fossil species hadseemingly disappeared, he would nothave been in a position to declareconfidently that he had found a massextinction level. First, there was no wayof estimating whether the rock layerrepresented a week or a million years oftime. Second, he could not determinewhether the contact with the rock unit

lying above the rock layer in questionindicated a smooth transition – a shortspan of time – or a gap of a week or 10million years after deposition of thefossiliferous layer. Further, thedisappearance of 50 species couldindicate nothing: they might havesurvived into the time represented by thegap and lived on for centuries, dying outand being replaced in a normal, gradual,Lyellian, way. Equally, even if therewere no gap in deposition, thenineteenth-century geologist would havebeen entirely unable to demonstrate thatthe 50 extinct species had not simplybecome extinct locally and survivedhappily elsewhere. Without refineddating methods, it is hard to demonstrate

a global mass extinction.The fundamentals of stratigraphy had

been established by 1840, and thedetails were increasingly ironed outduring the later nineteenth century. TheSystems, Silurian, Devonian,Carboniferous etc, identified and namedby Murchison and his colleagues in the1830s, were subdivided into epochs,stages and zones, and these finer unitswere traced over large areas. In the caseof widespread marine sediments, forexample those of the Early and LateJurassic, this exercise proved to behighly successful. Using ammonites,individual zones, and even narrow beds,could sometimes be tracked acrossEurope, from the windy cliffs of Dorset,

in southern England, south to the JuraMountains in southeastern France, andwest across the wooded hills of southernGermany. However, until 1920, therewas no method to establish time spans.

In their minds, most Victoriangeologists had long abandoned the literalestimate, based on the Bible, of an Earththat was only 6000 years old. They had aconcept of the vastness of geologicaltime – the sheer thickness of the rocksproved that – and they were thinking ofmillions of years, but whether 5 millionor 5000 million, no one could say. Mostlate Victorian geologists settled on afigure of about 100 million years.However, such efforts at dating werepurely speculative, and most geologists

thought it pointless to indulge in suchwild discussions.

Precision in dating rocks came atlast after the discovery of radioactivityin the late nineteenth century by HenriBecquerel in Paris. The first use ofmeasurements of the timing ofradioactive decay – radiometric dating –came in 1906, following fundamentalwork by Ernest Rutherford. Radiometricdating, from the start, provided exact ageestimates in millions of years. Initialefforts set the broad scale – that thefossiliferous units of the Palaeozoic andMesozoic were hundreds of millions ofyears old, and that the Earth wasthousands of millions of years old.

Work since 1920 has increasingly

refined those age estimates, andpalaeontologists have been able toachieve some confidence in estimatingthe time represented by a single rockunit, or the probability of a gap.However, the Victorian problems arestill live issues, since radiometric datingcannot be applied to most rocks,especially fossil-bearing sediments, andthe precision is often poor. In recentyears, most geologists have accepted adate of between 245 and 250 millionyears ago for the Permo-Triassicboundary. Recent, more precise datingmethods, have yielded values around251 million years ago, fantasticallyaccurate in geological terms, but still notgood enough to determine, for example,

whether the crisis interval lasted for afew years or a million years.

Richard Owen and theextinction of his Dinosauria

But surely, despite these nigglingproblems, the Victorian palaeontologistsmust have been impressed by theextinction of the dinosaurs, just aseveryone is today? Dinosaur extinctionis such an icon that it cannot be ignored.Quibbling over precision of datingshould hardly affect the magnitude ofsuch an event.

Certainly, the great reptiles of the

Mesozoic, the marine ichthyosaurs andplesiosaurs, the flying pterosaurs and thedinosaurs, were well known by 1840from several dozen specimens. But, noone then seemed to have an inkling thattheir final disappearance might havebeen part of a mass extinction. This isreally no surprise when the facts areconsidered. In 1840, only 20 or 30species of these reptiles were known intotal, from a range of rocks in Europedating from the late Triassic to theCretaceous. There was no need topostulate such a thing as a massextinction. Each species of dinosaur orplesiosaur had existed for its span oftime, whether 10,000 or 10 millionyears, and it had then succumbed to

some local crisis or had been replacedby another species.

Richard Owen, who named theDinosauria in 1842, had some words tosay about the extinction of the giantMesozoic saurians. Indeed, hiscomments have sometimes been takenrather out of context and misinterpretedas perhaps the first published hypothesisabout dinosaurian extinction. Not so.Owen argued in 18423 that the Creatorhad chosen the Mesozoic era as suitablefor dinosaurs because of its differentatmospheric conditions. He believed thatthe air then was deficient in oxygen, andthat this suited the dinosaurs. Asreptiles, they had lower metabolic ratesthan the birds and mammals and could

survive on less energy. Owen arguedthat oxygen levels rose during theMesozoic and that the atmospherebecame more ‘invigorating’. The worldthen became uninhabitable for the hugesaurians, and they died out, together withthe giant marine reptiles and thepterosaurs.

This argument was essentiallycircular, since Owen’s evidence for lowoxygen levels in the Mesozoic wassimply the presence of dinosaurs andother prehistoric reptiles and the virtualabsence of mammals and birds. He wasexplaining the existence of these earlyreptiles in Mesozoic rocks in terms of aprecise matching by a benevolentCreator of living things and physical

environments. This is pure Lyell. Owenwas not primarily trying to explain whythe dinosaurs died out when they did –for him, that would have been apreordained event in the Creator’s plan.

Mr Darwin’s view

Darwinism largely put an end to suchodd mixtures of Christianity and Lyell,but Charles Darwin could not escape theproblems of dating and the quality of thefossil record. Despite his brilliantinsights in other evolutionary topics, hewas not able to make the conceptual leapto deny Lyell and to accept the reality ofmass extinctions.

Charles Darwin devoted twochapters of his On the origin of speciesby means of natural selection topalaeontology. He discusses the‘imperfection of the geological record’in chapter 9, and surveys the history oflife in chapter 10. Here, Darwinaddresses the issue of mass extinction,but tentatively:4

the utter extinction of a group is generally, aswe have seen, a slower process than itsproduction. With respect to the apparentlysudden extermination of whole families ororders, as of Trilobites at the close of thepalaeozoic period and of Ammonites at theclose of the secondary period, we mustremember what has already been said on theprobable wide intervals of time between ourconsecutive formations; and in these intervals

there may have been much slow extermination.

In other words, apparent massextinctions, such as the end-Permianevent (‘the close of the palaeozoicperiod’) were probably illusory, the sumof many minor local extinctions partlyhidden in a gap in the rock record.Darwin, of course, explained thatspecies naturally became extinct fromtime to time as a result of competitionwith other species and their replacementby superior competitors.

Darwinian palaeontologists of thelatter half of the nineteenth centuryfollowed the master in saying nothingabout dinosaur extinction. Darwin’sstrongest supporter, Thomas Henry

Huxley (1825–95), wrote a number ofarticles about dinosaurs, but neverdiscussed their extinction. For example,in one of his first papers on the subject,published in 1870, ‘On the classificationof the Dinosauria’, Huxley5 describedthe 16 species of dinosaurs known up tothat time from fossils of Triassic,Jurassic and Cretaceous age found inEurope, North America, Africa andAsia. Dinosaur extinction? Not a word.

Later Victorian accounts were nodifferent. Othniel Charles Marsh (1831–99) was in a strong position to talk aboutthe end-Cretaceous mass extinction ofthe dinosaurs, but he did not. Marsh wasa leading North American vertebratepalaeontologist, famous for his

involvement in the ‘bone wars’ of the1870s to the 1890s, when he and hisarch-rival, Edward Drinker Cope(1840–97), vied with each other tounearth and name as many dinosaurs aspossible from the American Midwest.Marsh wrote a number of reviews of thediversity of the dinosaurs, covering thesame ground as Huxley had, butincorporating also all the new NorthAmerican finds. In 1882, Marsh listed46 dinosaur genera, and in 1895 this hadrisen 68.6 In these papers, he showedhow the different dinosaur dynasties hadwaxed and waned through the Mesozoic,but their final departure from the worldwas not discussed.

But what of the other mass

extinctions that we now recognize?Virtually nothing was written about theend-Permian event and the others. Butthe most recent extinction event, whenlarge hairy mammals, such asmammoths, mastodons, woolly rhinosand others died out did attract repeatedattention in Victorian times.

Floods or glaciers?

The late Pleistocene extinctions, some10,000 years ago, seem to coincide withthe retreat of the last ice sheets fromnorthern Europe and North America. Ithad been clear to early geologists andpalaeontologists that strange, exotic

beasts had once lived in these parts ofthe world. As we saw in Chapter 1,specimens of mastodons from NorthAmerica, mammoths from Europe andother large hairy mammals were muchdiscussed in the eighteenth century.

By 1850, many more specimens hadbeen collected. William Buckland, forexample, who described the firstdinosaur, Megalosaurus, in 1824, wasat the same time engaged in studies thathe regarded as much more important. Hehad directed excavations at KirkdaleCavern in Yorkshire, which had beendiscovered in 1821. Buckland foundabundant bones of deer, hippos, rhinosand mammoths in the cave, together withhyaena bones and coprolites (fossilized

faeces). A zookeeper had drawnBuckland’s attention to the fact that thecoprolites contained crushed shards ofbone, and that they looked just like theexcrement of the modern hyaenas thatwere in his care. Buckland formulatedthe view that the cave had been a hyaenaden, and that those scavengers haddragged back carcasses, and parts ofcarcasses, to the cave. Here indeed wasa scene that differed considerably frommodern Yorkshire!

In his Reliquiae Diluvianae,published in 1822,7 Buckland arguedthat climates had been warmer inPleistocene times, hence explaining theexotic, rather African, fauna. Moreimportantly, though, he accounted for the

extinction (at least the local extinction)of these exotica by the universal Floodrecorded in the Bible. The rising watershad trapped the hyaenas in their caveand had killed off the large, exoticmammals throughout England, and alsoin the rest of Europe.

Buckland lived to give up his flood-based viewpoint. Increasing evidencewas found in the 1830s that northernEurope had been swept by vast icesheets. The shape of the landscape,especially in Scotland, Scandinavia andaround the Alps, showed that glaciershad gouged deep, smooth-sided valleys.In highland areas, great pavements ofbare rock were to be seen, still bearingscratch marks produced by ancient

glaciers. Erratic rocks lay everywhere –boulders that had been torn up bymoving glaciers and dumped miles fromtheir original source.

The glacial model was championedespecially by Louis Agassiz (1807–73),the famous Swiss geologist and experton fossil fishes. He argued that Europeanclimates had once been warm andequable – entirely suitable for largeAfrican and Asiatic mammals such ashippos and elephants. The cooling of theclimate, and the march of the ice, hadbrought all this to an end:8

The appearance of this great cover of ice musthave caused the extinction of all organic life onthe surface of the globe. The territory ofEurope, recently covered by tropical vegetation

and occupied by herds of elephants,hippopotamuses and gigantic carnivores, founditself entombed under a vast mantle of ice thatcovered fields, lakes, seas and plateaus alike.To the movement of a powerful creationsucceeded the silence of death. Springs driedup, streams ceased to flow, and the sun’s rays,in rising over those frozen expanses (if theystill reached there), were met only by thewhistling of the northern winds and by thethunder of crevasses as they split the surface ofthis vast ocean of ice.

Pleistocene overkill

Geologists of the time generallyaccepted Agassiz’s remarkable insightthat Europe and North America had been

covered by ice during the Pleistocene.But very few accepted his seeminglycatastrophic viewpoint that largeswathes of life were wiped out by thecold and the ice. Indeed, Charles Lyell,while admitting the truth of the ice age,based as it was on clear physicalevidence scattered through thelandscapes of Europe and NorthAmerica, was extremely uncomfortablewith the idea of a sudden intensefreezing and global extinction. He noted,correctly, that the large mammals hadlived on during the times of cold and thatthey had clearly been adapted to theconditions. Lyell’s view was that thevarious large Pleistocene mammals haddied out one by one, for a variety of

reasons, but that there had not been anysuch thing as a single extinction event.

For the remainder of the nineteenthcentury most geologists preferredLyell’s viewpoint. They did not identifya single extinction event, but linked theextinctions to changing climatesthroughout the northern hemisphere.Extinctions had been gradual, and hadbeen caused by particular changes inclimates and local conditions – perhapsdifferent causes for the extinction ofeach species.

A role for humans in these mostrecent extinctions had been discussedsince the first discoveries of mastodonsand mammoths. Could stone age peoplesin different parts of the world have

killed off the large mammals? After all,they would have made attractivepropositions for dinner. Lyell rejectedsuch a notion, since initially he wasconvinced that humans had not appearedunt i l after all the late Pleistocenemammalian extinctions had taken place.By 1860, however, archaeologicalevidence proved that humans andmammoths, and other large Pleistocenemammals, had co-existed – bones of thegiant mammals were found in closeassociation with human bones andartifacts. Lyell was forced to change hismind, and he admitted a possible rolefor humans in the demise of thePleistocene mammals. At the same time,a French investigator was coming up

with convincing evidence for a differentcatastrophic explanation for the events inthe late Pleistocene.

Boucher de Perthes: the lastcatastrophist?

Jacques Boucher de Perthes (1788–1868), a French civil servant, wasinstrumental in providing the evidencethat early humans and the Pleistocenemammals of Europe had cohabited. Heknew that in order to convince thedoubters, such as Lyell, he had to workwith scrupulous care. Boucher dePerthes excavated sites around

Abbeville in the valley of the Somme,and he reported stone tools from layersbelow bones of woolly mammoth,woolly rhinoceros and hippopotamus.He argued that his excavations haduncovered remains from before thebiblical deluge.9 The Flood had thenswept over the world, destroying thestrange mammoths and rhinos in Europe,together with the primitive humans thathunted them. After the Flood, new,modern, animals filled up the territoriesof Europe. This idea of a catastrophicflood came directly from GeorgesCuvier in the 1820s. In France andGermany such catastrophist viewpointsstill held sway – Lyell had not been ableto convince everyone.

British palaeontologists, however,were unable to accept a revival ofcatastrophism. Lyell himself argued thatsuch a huge catastrophe could not beproposed since so many of thePleistocene mammals still survivedtoday – rats, mice, shrews, foxes,wolves. He noted that it was only thelarger mammals that had been wiped out.In the 1860s, Richard Owen supportedLyell’s viewpoint, and he went further inpointing to what he thought wasconvincing evidence for the overkillhypothesis. Based on his work inAustralia, Owen argued that many of thelarge marsupials – giant kangaroos, hugewombat-like herbivores and others –had disappeared only when the first

humans arrived in Australia. In NewZealand, where he studied the giantflightless birds, the moas, Owen foundeven more convincing evidence that theMaoris had killed off these birds in thepast hundreds of years.

The great biogeographer andnaturalist, Alfred Russel Wallace(1823–1913), was also convinced by thearguments of Lyell and Owen. He couldsee the great threats that humanexpansion posed for the wildlife of thetropical world in his day, and hence hefound it plausible that earlier phases ofhuman migration had killed off nativeplants and animals.

In late Victorian times, then, severalinfluential commentators in England –

Lyell, Owen, Wallace – were strongsupporters of the overkill hypothesis.The latest of the documented extinctionevents, the death of large mammals onmany continents in the late Pleistocene,had indeed been rapid. But Boucher dePerthes’ idea that the rapidity implied asudden global catastrophe was clearlyunacceptable. Human activity provided aconvincing explanation: the timings ofhuman migrations seemed to tie inclosely with the timings of theextinctions.

The overkill model also provided aclear way to maintain a gradualistic,uniformitarian view of the history of life.If the last of the great extinctions couldbe explained as a special case, a one-off

caused by humans, then there was noneed to invoke catastrophic explanationsfor earlier mass extinctions whenhumans were absent. But how did lateVictorian and early twentieth-centuryscientists explain the extinction of thePermian reptiles, and of the dinosaurs,two of the other well-known greatdyings of the past?

Directed evolution

In the late nineteenth and early twentiethcenturies, many palaeontologists andbiologists took up non-Darwinianviewpoints, with grand names such as‘orthogenesis’ and ‘finalism’, both of

which implied some kind of preordainedplan to evolution.10 These modelsassumed that evolution was directed insome way and that it proceededaccording to regular patterns. ThePermian synapsids, or the dinosaurs,could be viewed as primitive, lumberingbeasts that had to give way to moreadvanced forms. Mass extinctions – thereplacement of Permian plants andanimals by those of the Triassic – justhappened. It was part of the plan, anddidn’t really require an explanation.

This reversal in opinions aboutevolution may seem rather strange.Surely, after Darwin’s pronouncementsin the Origin of Species in 1859, and thedebates that surrounded them, scientists

had essentially accepted the correctnessof his views? However, it seems thatmost biologists in the late nineteenthcentury, although they claimed to beDarwinians, weren’t really. Theyaccepted evolution – that organismshave changed through time, and that theyare linked through lines of descent thatform a huge branching tree – but notmany could accept the seemingpurposelessness of Darwin’s evolution.

In general, though, the extinction ofthe dinosaurs remained a non-question atthis time. Standard textbooks of generaland vertebrate palaeontology of thelatter half of the nineteenth century andthe first decades of the twentieth barelymention the subject at all. And if they do,

the explanation is brief. The dinosaurssimply came and went in their due time.As an example of this view, ArthurSmith Woodward (1864–1944), chief ofpalaeontology at the Natural HistoryMuseum in London, and later to befamous for having been duped intodescribing the famous forgery, Piltdownman, as a new species of hominid, wrotein his 1898 textbook11 that ‘toward theclose of the Mesozoic period … theDinosaurs gradually became extinct’.And later he states that the dinosaurs ofthe Cretaceous ‘became morespecialized and almost fantastic justbefore they disappear’.

This low-key approach to theextinction of the dinosaurs continued

remarkably late in the twentieth century.I scanned a dozen standard textbooks,some of them published in England,others in the United States and inGermany, dating from 1902 to 1968, andthere is barely a mention of the greatend-Cretaceous mass extinction event.This absence of debate in the textbooksprobably reflected the general opinion ofvertebrate palaeontologists. None theless, during the first half of the twentiethcentury, debate about the demise of thedinosaurs did continue, and writersconcentrated on the ‘excess spinescence’(growth of spines) in the ancientreptiles.

Racial senility

When I was young I remember beingpuzzled by newspaper headlines thatsaid such things as, ‘British tradesunions are dinosaurs’, ‘Party leader isBrontosaurus, doomed to extinction’.As a mad dinosaur fanatic at the age ofeight, I knew that dinosaurs had beenvigorous and successful for over 150million years, so why did adultsconsistently use dinosaurs as a metaphorfor redundancy and inefficiency?

The popular metaphor stemmed fromscientific views of some seventy ormore years earlier. The orthogenesis andfinalism of the turn of the twentiethcentury led to views of racial senility –

the belief that certain long-lived groupsof animals became old and their store ofevolutionary novelty dried up. Therewas a parallel here between the life-span of an individual plant or animal andthat of an evolutionary stock. Youth andearly racial vigour were equated, andseen to be just as inevitable as the oldage and death of an individual and theracial senescence and eventualextinction of a major group. Accordingto this view, the dinosaurs had beenaround for a long time, and they simplyran out of adaptability. The remarkablehorns, frills and spines of some lateCretaceous dinosaurs were occasionallycited as evidence for this racial senility.

A typical early account was given by

Arthur Smith Woodward in 1909 in hisaddress to the British Association for theAdvancement of Science.12 He pointedto the great spinescence, excess growthand loss of teeth of the later dinosaurs asevidence. At the same time, theAmerican palaeontologist FrederickLoomis expressed similar views whenhe wrote about the bony plates along theback of Stegosaurus:13 ‘with such anexcessive load of bony weight entailinga drain on vitality, it is little wonder thatthe family was short-lived.’

These arguments are easy to refute.For example, Stegosaurus lived duringthe Late Jurassic, some 90 million yearsbefore the extinction of the dinosaurs.Likewise, there is no evidence that the

last dinosaurs, from the Late Cretaceous,were any more spinose, crested ortoothless than their forebears. None theless, expressions of pure racialsenescence are to be found in thewritings of many distinguishedgeologists and palaeontologists throughthe 1920s and 1930s, even though theirviews were challenged at the time.

In the end, such ideas of orthogenesisand the racial senility of the dinosaurswere effectively demolished by theadvent of the modern synthesis or neo-Darwinian model of evolution in the1930s and 1940s. This was a revolutionbrought about by the arguments of awhole host of brilliant youngevolutionists – Theodosius Dobzhansky,

Ernst Mayr, George Gaylord Simpson,Julian Huxley – who saw thatorthogenesis, and related ideas, werepure mysticism. They pulled in evidencefrom the new laboratory science ofgenetics, and combined it with the kindof field observations that Darwin hadmade.

The modern synthesis position,which is the view held today,represented a return to pure Darwinism,with the addition of new sciences thatDarwin could never even have dreamedof. There was no longer any place forpreordained patterns in this view ofevolution. Of course, although thescientists rejected racial senility longago, it is hard for many people to shed

such views completely: surely we haveto believe in progress of some kind?Surely the dinosaurs were simplydoomed to extinction by their large size?

Baron Franz Nopcsa: spyand theorizer

During the 1920s and 1930s, a numberof authors eschewed racial senility, andconcentrated instead on biotic andphysical factors that might have causedthe extinction of the dinosaurs. BaronFranz Nopcsa (pronounced ‘NOP-sha’;Fig. 9) was one of the first. Indeed,Nopcsa was an original in many ways.

His full name was Baron Franz (orFerenc) Nopcsa von Felsö-Szilvás(1877–1933), hinting at his aristocraticorigins. He was in fact the last in a longline of noblemen with estates inTransylvania, lands that are now sharedbetween Romania and Hungary.

In 1895, Nopcsa’s sister, Ilona,found some huge bones at Hatçeg on thefamily estates. She showed them toFranz, who took them to Vienna, capitalof the Austro-Hungarian empire, foridentification. No one could help youngFranz, and in the end he decided to learnabout palaeontology and study themhimself. What Ilona Nopcsa had foundwere some of the last dinosaurs tosurvive in Europe, from rocks dating

from the very end of the Cretaceous.Thus began Nopcsa’s distinguished

career as a maverick, but brilliant,palaeontologist. He moved freelythroughout Europe, using his urbanityand his astonishing command oflanguages to attach himself to scientificsocieties in all countries: hispublications appeared in perfectGerman, English, French or Hungarian,according to his interests at that moment.At the same time, Nopcsa lived the lifeof a freebooter, offering his serviceswhen the First World War broke out tothe Austro-Hungarians: he suggested thathe could secure the friendship ofAlbania by becoming its king. Thisproposal was rejected, but Nopcsa acted

as a secret agent for the empire.Nopcsa was really the first to

explore reasons for the extinction of thedinosaurs in a serious way. Hesuggested, for example,14 that the greatamount of cartilage that he believed wasnecessary for growth to huge size‘perhaps … was one of the causes forthe rapid extinction of the Sauropoda’.Later, Nopcsa summarized a number ofviews on the extinction of the dinosaurs:their ‘low power of resistance’, theirhuge size, a shortage of food, or ‘areduction in their sexual functions’. Hefocused particularly on the supposed‘increase in function of the hypophysis’– this is the pituitary gland, located inthe head, which controls growth. Nopcsa

believed that some malfunction of thepituitary caused the very large body sizeof the dinosaurs. Secretions from thepituitary caused giantism, partly by theproduction of large masses of cartilageas precursors of bone, and partly by aform of acromegaly, or pathologicalexcess thickening and overgrowth oflimb bones and facial bones. Nopcsawrote that ‘the increase in weight of thelimbs in the dinosaurs recalls the eunuchcondition’. Quite where Nopcsa got hisinformation about eunuchs is unclear.

9 The brilliant Transylvanianpalaeontologist, Baron Franz vonNopcsa, an early dinosaurianpalaeobiologist.

After the war, Nopcsa lost hugeamounts of money, mainly becausevarious governments confiscated hisestates in the ensuing chaotic conditions.He was put in charge of the Hungarian

Geological Survey in 1925, and thissaved his finances for a while. But, in1929, he left in a rage, and embarked ona 5600-kilometre tour through Italy andsouthern Europe. He travelled in amotorcycle combination with his faithfulsecretary and lover, an Albanian calledBajazid, at his side. The two menreturned to Vienna, but, beset byfinancial problems and poor health,Nopcsa finally shot Bajazid, and thenhimself, in 1933.

The German school ofPaläobiologie

Most other palaeontologists of the earlytwentieth century were neither as exotic,nor as imaginative, as Franz Nopcsa.Mass extinctions, and the extinction ofthe dinosaurs in particular, were notdiscussed by many people at all. Ofthose who did consider the question,most preferred to focus on climaticchanges, rather than internal, biologicalcauses as Nopcsa had done.

As an example, the noted Americanexpert on fossil mammals, WilliamDiller Matthew, presented evidence in1921 for a model of dinosaurianextinction that involved gradualtopographic change and progressivereplacement by mammals.15 His study ofthe late Cretaceous and the succeeding

Palaeocene in North America suggestedthat there was extensive mountainbuilding and continental uplift. Thedinosaurs, which were adapted tolowland and marsh situations, weredisplaced, and the placental mammals,which were adapted to upland zones,moved in.

Other suggestions were that climaticcooling was the cause, or that diseaselevels had risen markedly in the lateCretaceous dinosaurs, or that earlymammals ate all the dinosaur eggs, orthat volcanic eruptions wereresponsible.16 In a sense, the debate washeating up (though not very much, as weshall see). But at least some geologistsand palaeontologists were identifying

the fact that something unusual hadhappened, and that it deserved anexplanation. Indeed, some, but not all, ofthese authors, recalled that it wasn’t justthe dinosaurs that died out 65 millionyears ago. Clearly, any satisfactoryexplanation had to take account of all theother victims on land and in the sea.

In 1929, a remarkable, but largelyforgotten, paper appeared in a Germanscientific journal calledPalaeobiologica. This journal was setup in 1928 to mark a new wave ofthinking in German palaeontology. Thediscipline was clearly namedpalaeobiology – not palaeontology – toset it apart from traditional approaches.Palaeobiologists were not simply

interested in fossils for the purpose ofdating rocks; they wanted to treat themas once-living organisms. This school ofthought hoped to take over the world ofpalaeontology with its new breed ofyoung scientists who used biological andbiomechanical approaches in studyingthe life of the past.

The paper, by Alexander Audova, anEstonian palaeobiologist, presented a61-page review of the whole question ofthe extinction of the dinosaurs.17 Audovarejected racial senility and simplenatural selection as explanations, andfocused on environmental change. Hisfavoured view, after surveyinggeological evidence onpalaeotemperatures and physiological

evidence on the thermoregulation ofmodern reptiles, was that temperatureshad declined gradually worldwide, andthat this acted directly on the dinosaursand other Mesozoic reptiles bypreventing proper embryonicdevelopment.

One hundred theories forthe death of the dinosaurs

In the time from 1920 to 1990, at leastone hundred theories for the extinction ofthe dinosaurs were proposed, althoughclearly at the low rate of only one or twoper year. No stone or fossil was left

unturned in the search for ideas, andthese ranged from the purely biological,to interactions among species, tochanges in environments, toextraterrestrial causes. It’s impossible tocatalogue in detail all the suggestionshere. In 1964, the American dinosaurexpert G. L. Jepsen was able to list 40.In 1990, I was able to identify over 100separate theories.18

I did not include casual remarksmade by palaeontologists at scientificmeetings, nor did I include any of theconstant flood of speculative pieces inthe newspapers. (Except the latest onefrom late 2000, which suggested that thedinosaurs had gassed themselves out ofexistence. A French palaeontologist had

apparently calculated that a cowproduces enough gas each week to fill abarrage balloon. A dinosaur weighingfifty times as much as a cow wouldtherefore produce fifty times as muchgas. Digestive gas is mainly methane, sobillions of gallons of methane enteredthe atmosphere each year through thedigestive systems of the dinosaurs. Themethane replaced the oxygen in theatmosphere, and the dinosaurs wereasphyxiated. Oddly, this theory has notyet been published in a reputablescientific journal.)

My 1990 list does not includehearsay and student japes, only theoriesthat were seriously proposed through thenormal channels of scientific publication

– which means that the papers shouldhave been checked by at least two orthree experts before proceeding topublication.

One hundred theories for the extinctionof the dinosaurs, presented from 1842 to1990.

Biotic causes (26): Medical problems:slipped vertebral discs; malfunction orimbalance of hormone systems;overactivity of pituitary gland andexcessive (acromegalous) growth ofbones and cartilage; limb bones tooheavy; pathological thinning of eggshells; diminution of sexual activity;cataract blindness; disease (caries,arthritis, fractures, and infections);

epidemics; parasites; AIDS caused byincreasing promiscuity; change in theratio of DNA to cell nucleus. Mentaldisorders: dwindling brain andconsequent stupidity; absence ofconsciousness, and absence of the abilityto modify behaviour; development ofpsychotic suicidal factors;Palaeoweltschmerz. Genetic disorders:excessive mutation rate induced by highlevels of cosmic rays; abortion ofembryos by cosmic rays.

Racial senility (6): evolutionary driftinto senescent overspecialization, asevinced in gigantism, spinescence orexcess armour; racial old age (WillCuppy:19 ‘the Age of Reptiles endedbecause it had gone on long enough and itwas all a mistake in the first place’);increasing levels of hormone imbalance

leading to ever-increasing growth ofunnecessary horns and frills; head tooheavy to lift.

Biotic interactions (6): competitionwith mammals; competition withcaterpillars which ate all the plants;overkill capacity by predators (thecarnosaurs ate themselves out ofexistence); egg-eating by mammals;consumption of all plants by giantdinosaurs; methane poisoning fromdinosaur flatulence.

Floral changes (11): spread ofangiosperms and reduction in availabilityof gymnosperms and ferns, which led toa reduction of fern oils in dinosaur diets,and to lingering death by terminalconstipation; loss of marsh vegetation;increase in forestation, leading to a loss

of habitat; reduction in availability ofplant food as a whole; presence ofpoisonous tannins and alkaloids in theangiosperms; presence of other poisonsin plants; lack of calcium and othernecessary minerals in plants; rise ofangiosperms, and of their pollen, led toextinction of dinosaurs by terminal hayfever.

Climatic change (12): climate becametoo hot (high temperature inhibitedspermatogenesis, unbalanced themale:female ratio of hatchlings, killedoff juveniles, or led to overheating insummer); climate became too cold (toocold for embryonic development,dinosaurs too large to hibernate, froze todeath in winter); climate became too dry;climate became too wet; reduction inclimatic equability and increase in

seasonality.

Atmospheric change (7): changes in thepressure or composition of theatmosphere; high levels of atmosphericoxygen leading to fires; low levels ofcarbon dioxide removed the ‘breathingstimulus’; high levels of carbon dioxidein the atmosphere and asphyxiation ofdinosaur embryos in the eggs; extensivevulcanism which produced volcanic dust;selenium, or other toxic substances,which caused thinning of dinosaur eggshells.

Oceanic and topographic change (12):sea levels rose; sea levels fell; floods;mountain building; drainage of swampand lake habitats; stagnant oceans causedby high levels of carbon dioxide; bottom-water anoxia; spillover of Arctic water

(fresh) from its formerly enclosedcondition into the oceans, which led toreduced temperatures worldwide,reduced precipitation, and a 10-yeardrought; reduced topographic relief, andreduction in terrestrial habitats; break-upof supercontinents.

Other terrestrial catastrophes (5):sudden vulcanism; fluctuation ofgravitational constants; shift of theearth’s rotational pole; extraction of themoon from the Pacific Ocean; poisoningby uranium sucked up from the soil.

Extraterrestrial explanations (15):entropy (increasing chaos in theUniverse and hence loss of largeorganized life forms); sunspots; cosmicand ultraviolet radiation; destruction ofthe ozone layer by solar flares and entry

of ultraviolet radiation; ionizingradiation; electromagnetic radiation andcosmic rays from the explosion of anearby supernova; interstellar dust cloud;flash heating of atmosphere by entry ofmeteorite; oscillations about the galacticplane; impact of an asteroid; impact of acomet; comet showers.

Problems with the‘dilettante’ approach

Why focus here on the extinction of thedinosaurs? After all, the topic of thisbook is the end-Permian event, some185 million years earlier. However,

very little was ever said, or written,about the end-Permian mass extinctionuntil recently. Despite the fact that theend-Permian event was so much vasterthan the end-Cretaceous crisis, allattention focused on the demise of thedinosaurs. And this was mostunfortunate, since it lent an air ofamateurism to the whole topic of massextinctions.

According to the varied list ofreasons for dinosaurian extinction (seeBox), anything is possible. Climatesbecame too wet, too dry, too hot, toocold: you can take your pick. Evidently,something is wrong here, and this feelingof a subject that was running out ofcontrol convinced many serious-minded

geologists and palaeontologists in the1960s and 1970s that they should keepclear of anything to do with massextinctions, not just the extinction of thedinosaurs. I have labelled this phase ofspeculation the ‘dilettante’ approach,meaning that it was beset by dabblers,by people who thought about the topicfor a while, published their pet idea, andthen moved on to another subject.

Certain of the suggestions in the listare indeed perfectly reasonable ideas onthe basis of present knowledge, but theobviously ludicrous nature of many hadimportant consequences. While seriousgeologists were quietly appalled by therandom speculation, others, often outsidethe direct fields of expertise required,

thought that mass extinctions, andparticularly the extinction of thedinosaurs, were fun, speculative topicsin which anyone could join.

Many of the ideas listed werepresented by non-palaeontologists, andcertainly most of the authors had littlefirst-hand knowledge of the lateCretaceous fossil record of dinosaurs –hence the ‘dilettante’ sobriquet. A largenumber of the theories, all of whichwere published in standard scientificjournals by scientists who were no doubtexpert in their own fields, show aremarkable relaxation of scientificstandards. It was as if, at the meremention of ‘dinosaur extinction’,scientists breathed a sigh of relief and

felt freed from the straitjacket of normalscientific hypothesis-testing.

It is important to remember that theextinction of the dinosaurs 65 millionyears ago was part of a larger massextinction, during which numerous othermarine and terrestrial groups died out. Inthe sea, the marine reptiles, particularlythe plesiosaurs and mosasaurs,disappeared, as did some major molluscgroups, the free-swimming ammonitesand belemnites, and the bottom-livingrudists (large, reef-building molluscsthat were fixed to the sea floor). Evenmore striking was the loss of hugediversities of microscopic plankton, inparticular the foraminiferans, tinyshelled protozoans. On land, of course,

the dinosaurs died out, but so too did theflying pterosaurs, as well as somegroups of birds and mammals. Thewhole mass extinction is generallytermed the KT event (K for kreta, Greekfor ‘chalk’, a common rock in theCretaceous, and T for Tertiary, thesubsequent geological period).

I believe that there are four mainarguments in support of the view that thestudy of the KT event had in fact run outof control by the 1960s and 1970s.

• Many of the authors demonstrated anignorance of basic palaeontologicaldata. For example, the hypotheses wereoften restricted to explaining why thedinosaurs alone died out, and no mention

was made of the marine plankton andother animals that also disappeared. Thequestion of the survivors of the KT eventwas often not tackled: some scenarioswere so extreme or catastrophic that it ishard to understand how the land plants,insects, frogs, lizards, snakes,crocodiles, turtles and so on were notdetectably affected. In other cases, thetiming of evolutionary events is wrong:for example, the flowering plantsappeared 40–50 million years before theKT event, the mammals 150 millionyears before. Neither group could havecaused the demise of the dinosaursunless some other major evolutionaryinnovation in one or the other isproposed.

• A number of the theories apparentlyignored basic biological principles.Could caterpillars really compete withherbivorous dinosaurs and eat all theplants? Could dinosaurs really havebeen like automata, and unable to modifytheir behaviour? Is it possible to modela terrestrial biosphere in which a singlefactor – epidemics, parasites, glandularmalfunction, competition, predation –would lead to a complete ecologicalbreakdown?

• The mode of argumentation in manypapers was by strong advocacy – thetechnique of course which Charles Lyellhad used to such effect in the 1830swhen he removed catastrophism from the

field of reasonable scientific discourse.An argument by advocacy runs like this :‘If it is assumed that dinosaurs wereendothermic/that UV radiation wasincreasing during the Cretaceous/thatcaterpillars competed for food withplant-eating dinosaurs, then it followsthat.… If it is further assumed thatclimates were becoming warmer, orcolder, or drier, or wetter, then itfollows that.…’ It is rare to find carefulweighing of evidence both for andagainst particular hypotheses.

• There is also the assumption by someauthors that the whole subject is reallyjust a parlour game, and not terriblyserious. If a dinosaur palaeontologist

were to write an account of his or hertheory of the origin of the universe or ofa cure for cancer or of why caterpillarsturn into butterflies, he or she wouldprobably fail to get into print in areputable scientific journal. However,most of the dilettante theories of theextinction of the dinosaurs werepublished in very reputable journals:Science, Nature, American Naturalist,Journal of Paleontology, Evolution,and so on. How did these speculators getaway with it?

Otto H. Schindewolf: crazytheorist or visionary?

In the mid-twentieth century, at a timewhen most English-speakingpalaeontologists steered well clear ofany talk of mass extinction, a powerfulthinker in Germany challenged theirtimidity head-on. Otto H. Schindewolf(1896–1971; Fig. 10) had a long careerby any standards, publishing his firstpaper in 1916, aged twenty, andcontinuing to publish his ideas onpalaeontology and stratigraphy until1970, a span of 54 years of professionalactivity. After the Second World War,he became professor of palaeontology atthe University of Tübingen, and hedominated his staff and students in theclassic German way: what he said waslaw, and could not be debated or

discussed. And yet, what he was sayingthen was completely at odds with whatwas being said in the English-speakingworld.

At the time when the leadingAmerican and British scientists wereestablishing the modern synthesis,linking Darwin’s classic writings withthe new discoveries in genetics, ecologyand palaeontology, Schindewolfremained all his life an anti-Darwinian.He had been influenced as a young manby older German ideas that species, andlarger groups, such as the dinosaurs, theammonites, the trilobites, passed througha ‘lifespan’ akin to the lifespan of anindividual. The group began with aninitial phase of explosive evolution (=

youth), followed by a period of stability(= middle age) and ended withdegeneration and extinction (= old ageand death). This concept received thegrand title of typostrophism.

10 Otto Schindewolf, doyen ofGerman palaeontology from the1940s to the 1970s, and promoterof neocatastrophism.

Whatever it was called,typostrophism was simply a version ofthe early twentieth-century idea of racialsenility that had been roundly rejectedby the neo-Darwinians. Such a pre-programmed history could have no placein the world of Darwinian evolution,since there is no genetic mechanism torecord and promote a ready-madehistory for a group.

There have been many differentviews of Schindewolf’s role. It has beensuggested that he virtually single-

handedly held back the development ofmodern evolutionary theory in Germanpalaeobiology, and even after his deathit was considered sacrilegious tocriticize his ideas. However, despite theserious anachronism of typostrophism,Schindewolf was also something of alone voice speaking up for massextinctions at a time when there was adeathly hush on the subject in theEnglish-speaking world.

In his most influential text,Grundfragen der Paläontologie,published in 1950 and essentially theBible for German palaeontologists fordecades, Schindewolf gave his view ofthe end-Permian events:20

The Permian system constitutes the end of thePalaeozoic, the age of ancient animals, and infact, there is at that point a break of majorimportance in faunal evolution. In the Permianwe find the last of the trilobites, which were sothoroughly characteristic of the Palaeozoic.Large groups of hydrozoans, brachiopods,crinoids and bryozoans of the old stamp dieout.… Several of these ancient groups [ofamphibians and reptiles] died out in thePermian and were replaced in the Triassicsystem by numerous new forms. In short, weencounter almost everywhere a radical contrastbetween the old and the new.

In his book, Schindewolf did not discussmass extinctions in any further detail –but that was soon to come.

In the 1950s, Schindewolfdeveloped his idea that extinctions, as

well as the other phases of thetypostrophic cycle, were not affected byphysical processes such as climaticchange, sea-level change, vulcanism orthe like. Instead, extinction was part ofthe evolutionary cycle, and its causeswere innate to the organisms. But, toexplain mass extinctions, when manydifferent typostrophic cycles seemed tocome to an end simultaneously,Schindewolf argued that cosmicradiation following the explosion ofsupernovas was the cause. The suddenbursts of cosmic radiation, he argued,led to an increase in mutation rate withinmany groups of organisms. This forcedthem into the declining phase of thetypostrophic cycle, when the runaway

mutations caused overspecializations,deleterious organs and finally extinction.

To supplement the development ofhis ideas about mass extinctionsSchindewolf embarked on a series ofstudies of the Permo-Triassic boundary.This programme of work was, inretrospect, highly innovative. He and hisstudents tracked down high-quality rocksections around the world that traversedthe boundary. He concentrated inparticular on the Salt Range successionsin Pakistan, and documented in detailhow the faunas changed. Schindewolfbegan to pull together the evidence fromdifferent sections, to produce global-scale documentation of the magnitude ofthe end-Permian mass extinction. But he

did not convince many geologistsoutside Germany.

Neokatastrophismus?

In a provocatively titled publication in1963 – ‘Neokatastrophismus?’ –Schindewolf claimed a place forcatastrophism in geology. In this paper,as in others at the time, Schindewolfargued with his critics, some of whomwere rightly sceptical about his theory ofcosmic rays, and others of whom evendenied that there had been an extinctionat the end of the Permian.

This denial of the end-Permian event

was particularly prevalent amongvertebrate palaeontologists. Twodistinguished experts on fossilamphibians and reptiles, the AmericanCharles L. Camp and the senior Britishpalaeontologist D. M. S. Watson, bothargued in the 1950s that the apparentchangeover among dominant vertebratesat the Permo-Triassic boundary wasprobably more to do with theimperfections of the fossil record thananything else. Camp suggested that thePermian amphibians and reptiles thatapparently disappeared might just havelived in habitats that were not preservedin the Triassic. Watson, similarly,suggested that, for vertebrates at least,the apparent end-Permian mass

extinction might be little more than a gapin the record.

Schindewolf was incensed by whathe regarded as such woolly thinking.21

He argued strongly that there were nogaps – the rock successions in the SaltRange, and in South Africa and Russia,where the amphibians and reptiles havebeen most discussed, were continuous,with no obvious breaks. There truly hadbeen a mass dying.

In the introduction to his paper,Schindewolf noted, perhaps a littlecoyly, that he had been termed ‘the mostimportant and most consequentialspokesman for the idea ofneocatastrophism in currentpalaeontology’. Indeed, his espousal of

mass extinction, and of cosmic rays fromexploding supernovas, was reminiscentof the dreaded catastrophism that Lyellhad so successfully disposed of in the1830s (although only in England – not inGermany). Schindewolf compared hisstandpoint with that of Cuvier, and hedefended the term neocatastrophism, asopposed to simply catastrophism, in thatCuvier and his supporters had operatedat an early stage in the development ofgeology as a science, and that theirviews had of course taken no account ofevolution.

But Schindewolf went further. Yes,Darwin had written about the extinctionof species and genera, and suchextinctions were clearly a normal part of

Darwinian evolution. As new speciesarise, older ones die out for a variety ofreasons. But, Schindewolf noted, over100 years had passed since Darwinwrote his classic work in 1859, andgeology and biology had moved on – weshould no longer feel bound to acceptwhat Darwin had said on these subjects.

This kind of anti-Darwinian remarkdid not promote the acceptance ofSchindewolf’s ideas by palaeontologistsin the English-speaking tradition, nor byGerman biologists. After all, thediscovery of the structure of DNA in1953, and the rapidly developingscience of molecular biology, wereconfirming Darwinism and the modernsynthesis time and time again.

Schindewolf made himself something ofan outcast, except within his own milieu.In doing so, he seemed to confirm toothers that catastrophism, whether it wasrestyled as neocatastrophism or not, wasstill wild and dangerous speculation.Lyell had surely been correct. But, ininsisting on the reality, and the greatmagnitude, of the end-Permian event, itturns out that Schindewolf was right, andhis critics were wrong.

The end of catastrophism …or not?

Charles Lyell’s rejection of

catastrophism in the 1830s, andDarwin’s espousal of gradualism inevolution in 1859, seemed self-evidentby 1900. True, catastrophism reared itsugly head from time to time, whether indiscussions of the extinction of thedinosaurs, or in Otto Schindewolf’swritings, but it was easy to ridicule suchideas. And ridiculed they were throughmuch of the twentieth century. Iremember discussing a paper written in1956 by the American palaeontologist,M. W. de Laubenfels, in which hesuggested that the dinosaurs had beenwiped out by a giant meteorite. One ofmy senior colleagues, a palaeontologist,said he had written to de Laubenfels atthe time to ask him why the turtles and

crocodiles had not also been wiped outby the great atmospheric disturbances.Had the meteorites, he asked, simplybounced off the turtles’ backs? Mycolleague received no answer to hisenquiry.

Nationality no doubt played a part.For a time, most of the neocatastrophistswere, like Schindewolf, German.American and British commentatorssimply chose to ignore their work: manyindeed could not read German, and themajority were probably largely unawareof the German literature since itappeared often in regional Germanspecialist journals that were not taken bymany libraries outside Germany. It waseasier, too, to think of the catastrophists

as crazy if they remained anonymous,and in a different country. But everythingchanged on 6 June 1980. De Laubenfelswas dead by then, but he had been right.The Earth had been hit by a giantasteroid 65 million years ago, and thatimpact did kill off the dinosaurs. Thismarked the beginning of a new era ofserious research into mass extinctions.

5

IMPACT!

The modern era of mass extinctionstudies began in 1980, with thepublication of the proposal that thedinosaurs had been wiped out by theimpact of a huge meteorite, or asteroid,on the Earth. The paper finally re-established catastrophism at the core ofgeology. This was one of the most daringpapers ever published, it was wide open

to refutation, it had immensely highheuristic value, it raised a storm ofprotest, and it was one of the mostinfluential publications in earth sciencesin the twentieth century.

The paper1 was titled‘Extraterrestrial cause for theCretaceous-Tertiary extinction’, and itappeared in the leading Americanweekly journal Science, in the issuedated 6 June 1980. The first author of thepaper was Luis W. Alvarez, who hadwon the Nobel prize for physics in 1968for his work in identifying subatomicparticles. The other authors were his sonWalter, a professor of geology, and theircolleagues Frank Asaro and Helen V.Michel, all at the University of

California at Berkeley.The paper was daring because the

authors had very little evidence to backup their very large claims. This in turnmeant that it was wide open torefutation: science does not work byproving cases – that is for lawyers.Scientific theories are the bestexplanations for a series ofobservations, but a counter-observationmight disprove the theory at any time.So, Luis Alvarez and his colleagues hadreally stuck their necks out, making largeclaims and predictions that could soeasily have been demolished. The paperwas heuristic, meaning that it opened upa whole array of new problems andpredictions, and essentially launched a

new branch of earth science studies, thetrue ‘neocatastrophism’, as opposed toOtto Schindewolf’s earlier brand of the1960s. The word heuristic is derivedfrom the Greek heuriskein, ‘to find’, asin eureka, ‘I have found it’.

The proposal that an impact hadkilled the dinosaurs offended many,palaeontologists in particular, since itcame from a physicist, and an explosionof controversy followed in the 1980s.2The paper was hugely influential since itpresented a hypothesis that could havebeen refuted readily – but it was not, andindeed new evidence arrived all the timethat bolstered it.

Turning the hypothesis onits head

The whole hypothesis presented by theBerkeley team turned on a reversal oftheir starting idea. Walter Alvarez andhis colleagues were seeking anindependent way to calculate rates ofsedimentation in ancient rock sequences.Geologists can easily measure thethicknesses of beds of sandstone,mudstone or limestone, but the thicknessis not really proportional to time: a thinband of mudstone might representhundreds of years of slow deposition offine particles in the deep ocean, while amassive bed of sandstone, 100 metres

thick, might have been dumped from acatastrophic slumping event in a matterof minutes or hours. How could achronometer be devised that told timeaccurately?

Luis and Walter Alvarez reasonedthat iridium might offer a solution.Iridium is a rare platinum-group metalthat occurs in only minute quantities onthe Earth’s surface – in fact, in parts perbillion. When the Earth formed, iridiumwas present, but it then segregated intothe core. Iridium is rare now on thesurface of the Earth since it arrivesessentially from extraterrestrial sources,in the fine rain of small meteorites,tektites and cosmic dust that settles overthe surface in a slow rain. If the rate of

arrival of iridium was known – say onemicrogram per square kilometre perhundred years – then the amounts indifferent thicknesses of sediment couldbe measured and the length of time eachbed of limestone, mudstone or whateverhad taken to be deposited could becalculated.

The problem, or one of manyproblems, was how to measure thequantities of such a rare element. Up tillthen there were no analytical machinesavailable that could even begin to detectsuch tiny quantities. It was here that LuisAlvarez’s knowledge of experimentalphysics came into play. He and hiscolleagues were able to build a neutronactivation machine that could achieve

the necessary levels of precision.The geologists decided to test their

new chronometer on some rock sectionsthat Walter Alvarez had been studyingnear Gubbio in north Italy. Here, in thevalleys to the east of the medievalwalled town, were great sequences,over 400 metres thick, of thinly beddedmudstones and limestones that happenedto span the Cretaceous-Tertiaryboundary. They seemed to representmillions of years of slow accumulationof sediment in a moderately deeptropical sea. The geologists sampledbed-by-bed, and brought the rockchippings back to California.

After painstaking treatment in thelaboratory in Berkeley, the team plotted

the values of iridium in the sedimentsamples. They expected to findvariations through time – low values inrapidly deposited beds, and highervalues in slowly deposited units – whichwould allow them to calculate therelationships between time and thicknessfor the first time. Normal concentrationsof iridium were very low, averaging 0.3parts per billion. But, at the KTboundary, they found a surprising result.Through a thickness of 10 millimetres,the iridium values shot up to 9 parts perbillion, a vast increase of 30 times the‘normal’ level (Fig. 11). How shouldthis be interpreted?

11 The iridium spike, as recordedby Alvarez and colleagues in 1980in the Gubbio section, northernItaly. In the Cretaceous-Tertiaryboundary interval, a thin clay bandcontains much higher values ofiridium than the units above andbelow (but note the variablevertical scale, in which the KTboundary unit is muchexaggerated). The discovery of thisiridium enhancement led to theproposal of a massive KT impact.

Following the protocol of theresearch programme, Alvarez and histeam should have reasoned that a 30-foldincrease in iridium concentrations meantthat the thin clay band at the KTboundary had simply taken 30 times as

long to be deposited as the clays andlimestones above and below. Such aconclusion would have been within thebounds of expectation, since it is clearthat sedimentation rates can varyenormously in the sea. However, theyturned the theory on its head, and madetheir daring prediction.

The iridium spike

Iridium levels that shot up to 30 timestheir normal level could mean that therock unit had taken 30 times as long asnormal to be deposited. Or, Alvarezthought, perhaps it meant that the arrivalof extraterrestrial material had suddenly

increased. In other words, the Earth hadbeen hit by an asteroid – a hugemeteorite.

The team cross-checked their resultsat another KT section, the Stevns Klintcliff section in Denmark. If they found asimilar iridium enhancement in theboundary clay they were vindicated. Ifnot, then the iridium spike at Gubbiowas merely a local phenomenon, andprobably indeed caused by unusuallylow sedimentation rates at that location.The Danish values came in: backgroundiridium abundances of 0.26 parts perbillion, pretty much the same as atGubbio, and then a spike in the boundaryclay of 42 parts per billion. So, atStevns Klint the spike represented an

enhancement of 160 times backgroundlevels, even more dramatic than atGubbio. That was enough – the teamrushed their results into print.

They calculated the size of theimpacting object by various means,based on the predicted volume ofmaterial dumped on the Earth, the size ofthe explosion required to send sufficientmaterial into the atmosphere so that itencircled the Earth, and the knownrelationship between impacting objectsand the energy they transmit. The theorywas based on observations of the effectsof huge volcanoes, such as Krakatoa, avolcano located in the Sunda Straitbetween Java and Sumatra, which haderupted in 1883.

The Krakatoa eruption shot anestimated 18 cubic kilometres of moltenrocks and ash into the air, of whichabout 4 cubic kilometres reached thestratosphere, the higher parts of theatmosphere, and remained there for twoyears or more. This fine dust encircledthe globe and the effects of the eruptioncould be detected as far away asEurope. Writers and artists in Europerecorded unusually brilliant sunsets forweeks after the eruption. So, reasonedthe Alvarez team, a major impact wouldproduce just the same effects, only moreso.

The equation

So far, we have managed to avoid anymathematics. Now is the time tointroduce the only equation in the book.The equation is so daring, and really sosimple, that it is a good idea to workthrough it. We are talking here of anenormously bold prediction from reallyquite minimal evidence.

The Alvarez team took their two thinboundary clays, from Italy and Denmark,and proposed to use them to support thenotion that a giant rock, 10 kilometres indiameter, had hit the Earth 65 millionyears ago, punching a vast hole in theatmosphere, slamming into the Earth’scrust, instantly vaporizing, excavating ahuge crater 100 to 150 kilometresacross, and flinging millions of tonnes of

rocks and dust into the atmosphere. Thedust encircled the globe, blacking out thesun for a year or more, thus preventingnormal photosynthesis in the plants, andhence cutting the base from food chainsin the sea and on land. Mass extinctionfollowed.

The calculation was based on theoccurrence of the 1-centimetre thickboundary layer in Italy and Denmark.The Alvarez team argued that this claywas no ordinary marine clay, derivedfrom normal sources. It was the ash ordust of the impact, material that had beenlofted into the stratosphere anddeposited over a few years, carryingwith it the iridium that marked itsextraterrestrial source. Their formula

was:

where M is the mass of the asteroid, s isthe surface density of iridium just afterthe time of the impact, A is the surfacearea of the Earth, f is the fractionalabundance of iridium in meteorites, and0.22 is the proportion of material fromKrakatoa that entered the stratosphere(in other words 4 divided by 18 cubickilometres). The surface density ofiridium at the KT boundary wasestimated as 8 × 10−9 grams per squarecentimetre, based on the local values atGubbio and Stevns Klint. Measurementsof modern meteorites gave a value for f

of 0.5 × 10−6.Running all these values in the

formula gave an asteroid weighing 34billion tonnes. The diameter of theasteroid was at least 7 kilometres. Othercalculations led to similar results, andthe Alvarez team fixed on the suggestionthat the impacting asteroid had been 10kilometres in diameter. This dimensionled to further simple calculations – butthere is no need here for any moreequations.

The relationship between impactingobjects and crater size is known.Observations of recent craters, andexperiments with massive cannon thatfire bullets and large steel balls into clayboards show that the crater is always 10

to 40 times the diameter of the impactingobject, 10 times for large impacts, 40times for smaller. So, a 10-kilometreasteroid excavates a crater 100 to 150kilometres across. Equally, the speeds ofasteroids and the energy they transmitare also known. The KT asteroidprobably entered the atmosphere at aspeed of 25 kilometres per second, andits energy was equivalent to 100 millionmegatonnes of TNT, perhaps 30 timesthe explosive power of all nuclearwarheads currently in arsenals aroundthe world today.

How did the asteroid kill all thedinosaurs? Obviously it would havekilled all life in the immediate vicinityof the impact, and probably for 1000

kilometres all around, because of theblast and burning that followed theimpact. The Alvarez group, followingearlier studies of the wider effects ofvolcanic eruptions, reasoned that a largeenough impact would fill the wholeupper atmosphere with dust, so blackingout the sun for more than a year.Absence of sunlight would preventphotosynthesis in green plants both onland and in the sea (the microscopicphytoplankton). If plants were unable togrow, plant-eating animals would die offand, in turn, so too would the flesh-eaters. Absence of sunlight would alsolead to freezing. Either way, much of lifeon land (Fig. 12) and in the sea woulddie out.

The scale of the postulated eventwas huge, matched by the audacity of theAlvarez group. Here, using the simplestof equations, they had reconstructed oneof the most dramatic and devastatingevents in the history of the Earth. Andbased on what? A couple of clay layersin Italy and Denmark. No wonder, in1980, their paper was met with intenseoutrage. Now, 20 years later, we knowthat they were largely right.Catastrophism is re-established at thecore of earth sciences. Cuvier andBuckland must be smiling in theirgraves, Murchison would probably begreatly relieved, while Lyell, his friend,would no doubt be squirming andseeking to find how he could

accommodate himself to the newevidence.

12 The impact and itsconsequences? Triceratops andTyrannosaurus rex, two of the lastdinosaurs, face-to-face with theirfate.

Dangerous catastrophists

Recall that Otto Schindewolf, the‘neokatastrophist’ of the 1950s and1960s, had been shunned except bypeople associated with his power basein Germany. Other catastrophists hadmet the same fate. For example, KenHsü, professor of geology at the SwissFederal Institute of Geology and aconvinced catastrophist, records3 how,

when he arrived in the United States in1948 to begin graduate studies, he waswarned away from such dangerousdoctrines. His supervisor, EdmundSpieker, told him to steer clear of thework of T. C. Chamberlin, who, at thebeginning of the twentieth century, hadproposed that sudden events –catastrophes – provided ideal markerhorizons over wide areas that could beused in correlating rocks. ‘He has donemore harm to geology than any person’with such ideas, declared Spieker.

At the same time, in the UnitedStates, geologists were debating theorigin of Meteor Crater in Arizona. Tous it looks like a crater, with its deepscooped centre, its raised rim and its

circular shape. Amazingly, influentialgeologists constantly opposed such aview. G. K. Gilbert, for example, theChief Geologist of the United StatesGeological Survey around 1900,consistently argued that this crater, andothers, had been produced by volcanicprocesses. Such a suggestion seems togo against all common sense, but suchwas the fear of catastrophism that mostpeople denied that the Earth had everbeen hit by a large meteorite.

The evidence for impacts had beenso obvious that it could be denied onlyby a supreme effort of will. MeteorCrater looks so like an impact crater thatmany geologists really had to twist andturn before they could accept G. K.

Gilbert’s pronouncements. Opinionsshifted relatively rapidly in the 1960s.Information was leaking out about thegreat Tunguska impact in 1908, when alarge meteorite had exploded over aremote part of Siberia, flattening treesfor miles around. But the real turningpoint came with Gene Shoemaker’sstudies of the Ries Crater in Germany.

The Ries crater

13 Map of the Ries crater in westBavaria, south Germany. The innerand outer rings of the crater areclearly visible, and the ejecta(suevite), material thrown up by theback-blast following the impact,can be mapped out (irregularzones inside and outside thecrater). Of course, much of theejecta has been eroded away orcovered up by later sediments.

Here was a structure in southernGermany, surrounding the medievalmarket city of Nördlingen, that lookedvery like a crater. Nördlingen stands off-centre in a vast circular structure,measuring 22–23 kilometres across (Fig.13). The bowl of the crater is shallowand filled with lake sediments that haveturned into rich soils for farming. As youdrive away from Nördlingen in anydirection, you come to a sharp incline.The road zig-zags its way up and thendown on to the surrounding plain whichlies several tens of metres higher thanthe lands round Nördlingen. This is thecrater rim. Giant rocks, some as large asa house, stand at crazy angles both insideand outside this zone.

Again, as with Meteor Crater inArizona, German geologists had tussledwith many explanations for the Riesstructure. Most argued for a volcanicorigin – that the structure had resultedfrom some kind of volcano that haderupted and then collapsed back onitself, leaving no more than a shallowbowl-like structure. The volcanic conehad sunk back into the Earth, or perhapsthe volcano had been explosive and hadnever built a cone. Interestingly, alltraces of volcanic ash and lava had alsosomehow disappeared. In 1911,Artillery Major W. Kranz carried outexperiments with gunpowder buried inmud. He believed he could show that anunderground explosion could have

produced the Ries structure. Somedoubters suggested the possibility of animpact, but their voices were ignored.

In 1960, Gene Shoemaker (1928–1997), the eminent astrogeologist whoworked for the United States GeologicalSurvey, visited Nördlingen. He laterbecame a household name with thediscovery with his wife of the asteroidShoemaker-Levy that crashed intoJupiter in 1994. He had read all he couldabout the Ries structure, and wasconvinced that it was a crater. It wasideal for his studies since it was notparticularly old – the sediments thatfilled the structure contained abundantfossils of molluscs, fishes and otheranimals and plants that gave a Miocene

age of some 14.7 million years ago.Shoemaker arrived at the suevite quarryat Otting, on the eastern crater rim, andset to work. He selected the suevitequarry as most likely to convince thedoubters. Suevite is a melt breccia madeup from a jumble of angular shards ofsedimentary rocks – limestones,mudstones and sandstones – set in aglassy matrix, which formed attemperatures as high as 600°C. The termsuevite came from the Latin Suevia, theancient Roman name for the region ofGermany called Swabia in English andSchwaben in German.

Shoemaker, and his colleague E. C.T. Chao from Washington, had alreadyexplored the mineralogy of rocks from

the bottom of Meteor Crater in 1959.There they had discovered an unusualform of quartz called coesite, whichformed only under very high pressures ofup to 300 kilobars. Quartz is thecommonest mineral in rocks on Earth.Only the high energy of a meteoritehitting the Earth’s crust was enough tomodify normal quartz to coesite, socoesite was a clear marker of impact.

Shoemaker arrived at the Ottingsuevite quarry on 27 July 1960 late inthe afternoon and, as he records:4

Quickly I took three suevite samples as thedarkness fell, and then we camped in a woodnearby. On the next day we drove toNördlingen, and I sent the samples to Chao inWashington. Within a few days they reached

him and quickly he had found coesite by X-raytests.

Shoemaker and Chao wrote up theirresults and published their paper,entitled ‘New evidence for the impactorigin of the Ries Basin, Bavaria,Germany’, in the Journal ofGeophysical Research in 1961. Theirevidence was so conclusive that therewas no serious opposition. Geologistshad been living a lie for so long –forbidding themselves to admit theobvious – that it was impossible to deny.If it looks like an impact crater andsmells like an impact crater, it is animpact crater.

Megablocks, suevite andmelt bombs

The paper by Shoemaker and Chao wasjust the beginning. Once geologists hadaccepted the obvious, they flocked to theRies crater to find out everything theycould about how large impacts work.Unlike older craters, many of theconsequences of the impact could stillbe traced in the geology of the areaaround Nördlingen and could thus beidentified.

The geometry of the crater and of thefallout rocks reveal what happened. Thetrue crater edge is clearly marked, and is22–23 kilometres in diameter (see Fig.

13). But there is a smaller inner ring,some 11–12 kilometres across. Thisresulted from bounce-back of the craterfloor immediately after the meteorite hadpunched deep into the crust. The ejecta –the rocks thrown out after the impact –extend up to 70 kilometres from themiddle of the crater. As would beexpected, the ejecta layer becomesthinner the further one goes from thecrater, and the mean size of the rocksalso diminishes – huge house-sizedboulders in and around the crater, andsmaller gravel-sized material 40kilometres away. Some finer-grainedmineral materials called moldavite havebeen identified 600 kilometres to theeast in the Czech Republic.

Six unique rock types in and aroundthe crater tell the story of the events.First are the megablocks, huge irregularboulders, up to 25 metres across,composed of the underlying sediments,which lie at all angles just inside thecrater rim. These were clearly torn outof the centre of the crater, thrown intothe air and then fell back into the crater.

Second, the bunte breccia(‘coloured breccia’) are multicolouredblocks composed of fragments of therocks from beneath the crater – red andyellow Triassic sandstones andmudstones, grey and black Jurassicmudstones and limestones – in irregularhand-sized chunks set randomly in fine-grained rock dust. The bunte breccia is

found around the crater edge and in thesurrounding area, and it was clearlymade from material that was gouged outby the impact, thrown up as looseblocks, and then fused into an irregularmixed breccia as it landed.

Third, the crystalline brecciasconsist of fragments of deeper rocks,granites and other igneous rocks that liesome 400 metres below the surface.These were formed as the meteoritedrilled deep below the sedimentaryrocks and threw the basement crystallinerocks up into the area. Because of theiroriginal depth, and the force of theirremoval from such low levels, theserocks show evidence of metamorphism,that is, melting and distortion produced

by high temperatures and pressures. Thecrystalline breccias fell back into thecrater, and some are found outside it.

Fourth, the suevite, the rock typenamed first from the Ries crater, is abreccia composed largely of basementrocks, granites and gneisses, with rarepieces of sedimentary rock. The rockswere heated and up to 70% of anysample has been turned into glass. Theseparate fragments have been fused, butthere are many holes, indicating that thesuevite contained a great deal of gasduring its formation. The suevite isfound in great abundance, both insideand outside the crater; it had evidentlybeen thrown into the air during theimpact and had then fallen back. The

good citizens of Nördlingen andsurrounding towns had used the sueviteas a major building stone since theMiddle Ages – its warmth and easyworkability made it ideal – though littledid they know of the origin of theunusual stone!

Fifth, the impact melt materials arenatural glasses. These were formed fromthe local rock that had been melted bythe impact and had then fused as auniform glass. The molten material isfrequently shaped into ‘bombs’, disc-shaped glass bodies that had flownthrough the air, solidifying as theycooled. The external shape of the bombs– something like a frisbee – shows howthey solidified in mid-flight, and flow

patterns inside the bombs confirm howthe molten glass flowed in the secondsbefore it solidified.

The sixth, and final, impact rockconsists of the distant ejecta, materialthat was flung out after the impact andtravelled some distance from the crater.Some blocks of the underlying sedimenthave been located as much as 70kilometres away from the impact site,south of the Danube. More controversialare bentonite clays which are foundclose to the sediment blocks: these maybe the product of fine-grained ash-likematerial that was blasted into theatmosphere and which may haveblanketed a wide area around the crater.Finally, the moldavite tektites, small

meteorite-like melt fragments, of theCzech Republic, mentioned earlier, mayalso have been thrown several hundredkilometres by the force of the blast.

Geologists have now spentthousands of hours anatomizing the Riescrater and its ejecta. Their maps,boreholes, measurements andgeochemical studies of the extraordinaryrocks in and around the crater haveproduced one of the most detailedstories of an event that took no more thana few seconds. What happens during animpact?

Anatomy of an impact

The Ries asteroid struck 14.7 millionyears ago, and was 500 to 700 metresacross – more than half a kilometre. Itpunched through the outer atmosphereand hit the Earth’s crust at a speed of 20to 60 kilometres per second. The energyreleased was 100 megatonnes,equivalent to the explosive power of250,000 Hiroshima bombs. One hundredand fifty cubic kilometres of rock werethrown out of the crater and the resultantblast killed every living thing for 500kilometres around. Dust clouds travelledround the globe, and would certainlyhave caused startling sunsetsworldwide, as the eruption of Krakatoadid in 1883. The effect of the Riesimpact was considerably greater,

however, and the dust may have beensufficient to black out the sun for a fewdays at least.

The story of the Ries impact can bedissected into six phases. The asteroidhurtles down from the sky, passingthrough the atmosphere in one to twoseconds (phase 1). It penetrates throughthe Earth’s crust to a depth of about 1kilometre, cutting through 700 metres ofTriassic and Jurassic sediments and 300metres of underlying crystallinebasement rocks beneath (phase 2). Atthis early phase, small particles ofmolten material, the moldavite tektites,are thrown out, and a huge shock waveis generated downwards and sideways.

At the instant of impact, the pressure

is about 5 megabar, equivalent to 5million times normal atmosphericpressure, and the temperature rises to20,000°C. In phase 3, the meteorite andthe surrounding rocks to a depth of 1kilometre are compressed to less than aquarter of their original volume in underone-fifth of a second, and they vaporizeexplosively. The shock wave front isgenerated in all directions, travelling ata speed of 20 to 30 kilometres persecond, but it fizzles out after a fewkilometres. Below the deepest point ofthe crater, different zones of elevatedpressure and temperature can beidentified, the pressure and temperaturediminishing away from the crater.

Then the ejection phase (4) begins,

just two seconds after impact. As isoften said, to every force there is anequal and opposite reaction. The craterfloor bounces back, flinging up hugeamounts of rocks, melt products, ash andgases from the vaporized meteorite andsurrounding rocks. The mass of rocksand gas is shot upwards and outwards,and the ejecta front forms a conicalshape. The circular cone of hot ejectaraces sideways at the speed of anexpress train, stripping the soil androcks as it goes, and killing everything.At first, the largest megablocks andbunte breccia rocks fall back in andaround the crater, piling up tothicknesses of over 100 metres aroundthe crater rim. The thickness of the

ejecta blanket, and the average size ofthe rocks in it, decrease away from thecrater, since the energy of the ejectafront also diminishes as it racesoutwards.

The large, essentially unmelted,blocks of sediment are followed rapidlyby the suevite and melt products,including glass bombs, which are alsodumped in and around the crater for adistance of up to 100 kilometres. Finer-grained ash travels further, perhaps up to500 kilometres. After the ejecta blankethas passed, the landscape is devastated– irregular lumps of rock scattered atrandom are cloaked in a fine coating ofwhite ash and burning stumps of treespenetrate the ash here and there.

The entire process is over in 10minutes (phase 5). The meteorite hasgone, the double crater rim has formed –the inner rim created by the bounce-backprocess, some 11 kilometres wide, andthe outer rim, formed by sinking aroundthe inner ring, about 23 kilometresacross – and the rocks and ash havemainly fallen from the sky. The finestdust probably remains in thestratosphere for much longer, affectingclimates for some years after the impact.

Over the past 14.7 million years(phase 6), the crater has been muchmodified. It filled with a lake in theyears following the impact, and thicklayers of lake muds and limestones builtup. These sediments contain fossils of

green algae, reeds, freshwater snails,fishes, and even tortoises, snakes, birds,hedgehogs, hamsters, bats, martens,muntjac deer and other animals thatlived around the lake.

The Ries crater is still one of thebest documented of all craters. But, afterthe epochal paper by Shoemaker andChao in 1961, geologists began to spotcraters all over the place. In fact, craterhunting became something of a minorindustry, although only for a limitedfraternity. At first, it was easy. Craterfanatics identified all the obviousmodest-sized and relatively recentcraters they could see. They thenscanned aerial and satellite photographsfor circular structures that were too big,

or too old and too eroded, to be seen onthe ground. Many more were located inthis way, some of them as much as 100kilometres across. But this new wave ofcraterology did not make the receptionof the Alvarez paper in 1980 any easier.Most geologists were stilluncomfortable with catastrophism, andcould accept only a very limited numberof craters and impacts on the Earth.

Obdurate opposition?

In retrospect, it would be easy tocharacterize the opponents of the KTimpact model presented by Luis Alvarezand his colleagues in 1980 as luddites or

reactionaries. The impact proposal wasin line with all the new research on theRies crater, and on all the other cratersthat had been identified by 1980. At thetime, admittedly, there was no largecrater of the correct age that could bepointed to as the smoking gun that killedthe dinosaurs. But that wasn’t a strongcriticism: in the course of 65 millionyears, even a crater 100 kilometresacross could have been covered over bysubsequent sedimentation, or it could lieunder the oceans. The opponents ofimpact, however, had another line ofargument.

The standard view in 1980 has beentermed the gradualist ecologicalsuccession model. It was promoted

especially by geologists andpalaeontologists such as Bill Clemens,Leigh Van Valen, Robert Sloan andothers, who had studied the lastdinosaurs and the early mammals of thesame period.5 This model proposes thatthe dinosaurs, and other fossil groupsthat became extinct, were in decline longbefore the KT boundary, perhaps for asmuch as 5 million years. The concept ofecological succession proposes that thetypical Late Cretaceous floras andfaunas gave way to new floras andfaunas over a long span of time – soslow that an observer would not haveseen what was happening. But, after 5million years, the dinosaur-dominatedcommunities had given way to mammal-

dominated communities.This gradual ecosystem evolution

model is largely based on theprogressive appearance of a mammalcommunity (the ProtungulatumCommunity) of distinctive Tertiaryaspect in the last 300,000 years of theCretaceous in Montana. As the mammalsincreased in abundance, the dinosaursapparently declined, until theydisappeared altogether. This gradualreplacement is explained in terms ofdiffuse competition between dinosaursand mammals set against a major changein habitats. The lush subtropicaldinosaur habitats were apparently givingway to cooler temperate forests whichfavoured the mammals.

There is some doubt now about thecorrectness of the dating of the differentcommunities involved in this particularexample from Montana: perhaps some ofthe mammal fossils occur in muchyounger river channels that cut downinto Late Cretaceous sediments. Thisuncertainty about dating, however, doesnot remove all evidence for gradualextinction.

The gradualist model has beenextended to cover all aspects of the KTevents. In the seas, planktonicextinctions took 10,000 years, andvarious groups were already decliningwell before the boundary. A variety ofsea-bottom dwellers and filter-feedersdied out, but sea-bottom predators and

detritus-feeders were little affected.Extinction patterns of many marinegroups show gradual declines throughoutthe Late Cretaceous. The gradualistsexplain the long-term patterns ofextinctions by pointing to coolingclimates and major changes in sea levelat the end of the Cretaceous.

Gradualists also argue that the factthat many groups did not go extinct at theKT boundary is hard to understand in theface of some of the devastatingcatastrophist scenarios. On land,placental mammals, lizards, snakes,crocodiles, tortoises, frogs and otherfreshwater organisms show little sign ofextinction, and the plant record revealsonly modest and gradual changes. So,

argue the critics of a simple impactmodel for extinction, how did all thesegroups that lived side-by-side with thedinosaurs survive if the Earth had beendevastated by global blackout, freezing,catastrophic tidal waves, fires, masspoisoning and other disasters?

Also, and perhaps rather flippantly,the critics asked how the dinosaurs,ammonites and planktonic beasts knewthat the meteorite was about to strike?After all, these creatures had all begunto decline tens of thousands, or evenmillions, of years before the impact.

Ugly disputes: physics

versus palaeontology

Was there something more to the furiousdebate that followed the Alvarez paperin 1980? Perhaps the protagonists werenot simply debating scientific evidence.The two main models for dinosaurianextinction are based on rather differentkinds of data: essentiallypalaeontological and stratigraphic forthe gradualist models, and mainlygeochemical and astrophysical for thecatastrophic models. This meant that itwas hard for the proponents of one viewto assess the evidence that supposedlyfavours the other view. There is,however, apparently a more fundamentalsource of potential conflict between

certain biologists and physicists, or‘soft’ scientists and ‘hard’ scientistsrespectively, as they are often termed.6

The initial publication by Alvarezand colleagues was greeted scepticallyby many palaeontologists and geologistswith long-term expertise on aspects ofthe KT boundary. No doubt they resentedthe intrusion into their subject by a groupof physicists, and Luis Alvarez’s lengthycatalogue7 of his team’s credentials (aphysicist, two nuclear chemists and ageologist) may not seem so unusual inview of this resentment: ‘suddenly Irealized that we combined in one groupa wide range of scientific capabilities,and that we could use these to shed somelight on what was really one of the

greatest mysteries in science – thesudden extinction of the dinosaurs.’

The crux of the dispute was outlinedby Robert Jastrow, an astronomer andscience writer, in a report in the popularmagazine Science Digest:8

Professor Alvarez was pulling rank on thepalaeontologists. Physicists sometimes do that;they feel they have a monopoly on clearthinking. There is a power in their use of mathand the precision of their measurements thattranscends the power of the softer sciences.

The very titles of the Alvarez paperscould be seen to exemplify this: the 1980paper is titled ‘Extraterrestrial cause forthe Cretaceous-Tertiary extinction –Experimental results and theoretical

implications’, while an overview from1983 is ‘Experimental evidence that anasteroid impact led to the extinction ofmany species 65 million years ago’.Leigh Van Valen, a gradualist criticcommented9 that ‘to call [the Alvarez]evidence “experimental” is misleadingpropaganda; it refers merely to the factthat some observations were made in thelaboratory rather than in the field, not toan active experimental test.’

Luis Alvarez is surprisinglyrevealing throughout his 1983 paper, andhe is dismissive of his critics:10

I think the first two points – that the asteroidhit, and that the impact triggered the extinctionof much of the life in the sea – are no longerdebatable points. Nearly everybody now

believes them. But there are always dissenters.I understand that there is even one famousAmerican geologist who does not yet believe inplate tectonics.… People have telephoned withfacts and figures to throw the theory intodisarray, and written articles with the sameintent, but in every case the theory haswithstood these challenges.

He later outlines the advantages ofphysics in comparison withpalaeontology:11 ‘The field of dataanalysis is one in which I have had a lotof experience’ and ‘In physics, we donot treat seriously theories with suchlow a priori probabilities’. He furtherwrites, ‘That is something that made mevery proud to be a physicist, because aphysicist can react instantaneously when

you give him some evidence thatdestroys a theory that he previously hadbelieved.… But that is not true in allbranches of science, as I am finding out.’Public utterances from Luis Alvarezabout his ‘opponents’ were frequentlymore critical than these examples, to thepoint of being libellous, as reported atthe time by the journalist MalcolmBrowne in the New York Times.12

On the other hand, much of thedistrust of the physicists by certainpalaeontologists has surely beenunfounded, as David Raup, apalaeontologist and a catastrophist,notes. He quotes at length a statement byRobert Bakker, a dinosaurpalaeontologist, first published in the

New York Times:13

The arrogance of those people is simplyunbelievable. They know next to nothing abouthow real animals evolve, live and becomeextinct. But despite their ignorance, thegeochemists feel that all you have to do iscrank up some fancy machine and you’verevolutionized science. The real reasons for thedinosaur extinctions have to do withtemperature and sea level changes, the spreadof diseases by migration and other complexevents. But the catastrophe people don’t seemto think such things matter. In effect, they’resaying this: ‘We high-tech people have all theanswers, and you paleontologists are justprimitive rock hounds’.

Here is a mixture of righteousindignation and real concern. Part of thedispute revolved around different ways

of working, plus there was a realexpression of the supposed peckingorder in science, in which physics isseen as better, more reliable, thanpalaeontology. But much of the clashbetween the palaeontologists and theAlvarez camp in the early 1980s couldclearly be traced back to Lyell, and thelong-standing distrust of catastrophismof any stripe.

Styles of argumentation

Elisabeth Clemens has analysed thenature of the debate about ‘asteroids anddinosaurs’, and she argues14 that there

are many non-scientific undercurrents,such as styles of argumentation and therole of professional and popularpublication. She wanted to investigatejust why the impact theory gained suchrapid acceptance by diverse groups ofscientists and by the public, but wasinitially rejected by most geologists andpalaeontologists who were actuallyclose to the question. Why the mismatch?

It is important to note first that thebroadly based research enterprise thathas developed around the question of theextinction of the dinosaurs – geologists,palaeontologists, chemists, physicists,astronomers – is not a single communityof scholars, all of whom have had thesame training. It is a body consisting of

several factions, each going in differentdirections, and with very littlecommunication between them. Clemenssuggests that the Alvarez theory gainedrapid notice and acceptance in manyquarters because catastrophism ingeology was becoming intellectuallyfashionable. We have seen howShoemaker’s work opened the door in1961 for the acceptance of thepossibility of meteorite impacts on theEarth, and how he, and a growing armyof geologists, had developed asophisticated new branch of impactgeology by 1980. But there was more.

There had already been precursorsof Alvarez. As we saw earlier, deLaubenfels had seriously suggested in

1956 that an impact had wiped out thedinosaurs. He was a respectedpalaeontologist, but still the idea wasfairly comprehensively ignored.However, a flurry of extraterrestrialmodels had been published in the 1970sand an influential current of opinion15

had advanced the case for a supernova65 million years ago. The argument wasthat an exploding star had blasted theEarth with cosmic radiation. Althoughthis idea did not gain many adherents, itrested on much of the same evidenceused by the Alvarez group (but not theiridium spike). Perhaps the supernovatheory of the 1970s paved the way forAlvarez and impact.

Clemens argues that it was the mode

of presentation of the Alvarez hypothesisthat won it such wide attention andacceptance: ‘In a sense, the problem ofthe K-T boundary was framed so as tobe amenable to the methods of particlephysics’. The bulk of the long 1980paper (14 pages in all) is confined to thegeological and physical evidence for animpact, and the physical results of theimpact. The discussion of its biologicalresults occupies only half a page. Thepaper is restricted then to a rathersimple astrophysical hypothesis whichcould be tested in many ways, while themore complex aspects of stratigraphicimprecision and complexity of theevolution of biological communities arelargely omitted. These issues had to be

taken on board later, however.In subsequent publications in 1984,

the Alvarez team16 allowed from 10,000to 100,000 years for the overall length oftime involved in the extinctions, and theynoted that ‘the paleontological recordthus bears witness to terminal-Cretaceous extinctions on two timescales: a slow decline unrelated to theimpact and a sharp truncationsynchronous with and probably causedby the impact.’ However, by 1984, thesimplicity of the ‘instant-extinction’model of 1980 had ensured its generalacceptance by many scientists. The latermodifications are rather ad hocqualifiers that tend to protect the impacttheory from refutation by stratigraphic or

palaeontological evidence.The style of the arguments on both

sides of the debate in the professionalscientific literature may have beenimportant. But the role of the media,responding to intense public interest inthe debate, may also have fed back intothe science.

The role of the professionaland popular press

According to Elisabeth Clemens, thenature of the professional and popularpress has largely shaped thedevelopment of models of dinosaurian

extinction since 1980. She points out thatthe 1980 Science article was twice aslong as such articles usually are, andwas published in a prominent position,at the start of the issue. Each week, whenthe leading science journals Science andNature appear, the publishers chooseone or two articles for press coverage,and those articles are then reportedworldwide. The other articles, some 20or 30 each week, are not so favoured,and go unreported. This is no reflectionon the quality of the other papers, just afact of life.

The intense press coverageinevitably has a feedback effect, creatinga flurry of excitement around one or twopublications. This in turn inevitably

affects scientists – who are merelyhuman, and read the newspapers like anyother citizen. This one article by theAlvarez team gained a very widereadership, particularly in the UnitedStates, whereas other articles thatpresented similar theories at the sametime17 were much less widely read.

Following the success of the Alvarezpaper, it has been alleged that pro-impact papers were much favoured bythe editorial board of Science, and theargument spilled over into thecommentary and review sections ofleading journals and into thenewspapers.18 Clemens suggests that thevery format of publication has had arestrictive effect, since most of the

debate has been carried on so far in thepages of Science and Nature, both ofwhich normally publish only very shortpapers of four or five pages in length,and both of which require papers to bereadily understandable to a wideaudience. It is easier to present a simple,clear view, such as the impact, sheargues, than to debate the imprecision ofmethods for dating rocks, or thecomplexity of biological communities.

Iridium, shocked quartz,glass beads and the fern

spike

After 1980, and the immense commotionsurrounding the Alvarez paper,geologists and palaeontologists pursuedmany lines of investigation into the KTevent. It became ever clearer that,although there may have beensociological, methodological andpresentational criticisms of the strategyand style of the Alvarez team, they hadbeen essentially right. It would havebeen so easy for the impact theory tocollapse at any point, but it did not. Infact, new evidence supported it. Mostimpressively, kinds of evidence that hadnot been predicted by Luis Alvarez andhis colleagues lent independentcorroboration. What had been one of themost daring hypotheses in the earth

sciences, founded on extremely slenderevidence, was vindicated.

First, during the 1980s, geologistssampled KT boundaries all over theworld – they found the clay layer nearlyeverywhere they looked, and it wasenriched in iridium. The mere fact of theoccurrence of the clay is telling; theiridium enrichment is even moreimpressive. The clay and the iridiumwere found in all environmental settings,from rocks that had been deposited deepin the oceans, in shallow seas, on land,in lakes and rivers – everywhere. And itwas a worldwide phenomenon – fromCanada to New Zealand, from Russia tothe South Atlantic. The clay layerproved that some agency had sent a

plume of dust into the stratosphere, andthat the dust or ash had fallen outuniformly, cloaking the whole Earth in awhite blanket at least a few millimetresthick. In addition, that ash had broughtwith it iridium, a marker of impact. Thiswas what the Alvarez team hadanticipated. But some discoveries wereunexpected.

In many KT sections, as a result ofintensely careful search, geologistsfound shocked quartz and glass beads.As we have seen, quartz is thecommonest mineral on Earth, and itnormally occurs as largish grains,sometimes with irregularities andinclusions of other minerals, but neverwith any regular internal structure. At the

KT boundary, in the clays, geologistsfound quartz grains bearing crisscrossinglines, indicating that high pressures hadbeen applied. Some grains had four orfive sets of regular parallel lines runningacross: the more sets of lines, the higherthe pressure. Shocked quartz can beproduced during high-pressure volcaniceruptions, but volcanic shocked quartzusually has only a few sets of lines. Sothe KT shocked quartz was independentconfirmation of the high pressures of animpact.

In the Caribbean region and in thesouthern United States, geologists alsofound tiny glass beads, less than 1millimetre across. These looked just likethe results of melting during a volcanic

eruption, but the KT beads could nothave been produced by a volcano. Whengeochemists analysed the glass, theyfound that it did not have the chemicalcomposition of a lava, but it matched thecomposition of limestone and salt beds.How could a volcano produce moltenlimestone? These glass beads mustindicate that the meteorite hit limestonesand salts, melted the underlying rocks,and flung glass beads back out of thecrater.

The fern spike was anotherunexpected finding. Palaeontologistsfound an abrupt shift in pollen ratios atsome KT boundaries, showing a suddenloss of the pollen of flowering plantsand their replacement by ferns, and then

a progressive return to normal floras.This fern spike is interpreted asindicating the aftermath of a catastrophicash fall: ferns recover first and colonizethe new surface, followed eventually bythe flowering plants, such as grasses,bushes and trees, after soils begin todevelop. This was illustrated after theeruption of Mount St Helens in thewestern United States in 1980. The blastof the eruption flattened the trees andkilled off the grass and flowers for milesaround. The landscape was shrouded inash. Within a year, however, the firstplant life had returned: ferns hadsomehow survived beneath the covering,and they began to uncurl through the ash.It took a few years for the other plants to

return.

The Chicxulub crater

There was no need to find the crater, buta crater would surely help to clinch theKT impact model. After several falseleads, the crater that killed the dinosaurswas identified in 1991 on the Yucatánpeninsula in southern Mexico, centredaround the village of Chicxulub (Fig.14). The Chicxulub crater was buriedbeneath Tertiary sediments and could notbe seen at the surface, but boreholes andgeophysical evidence suggested that itwas 150 kilometres across – just the sizeLuis Alvarez had predicted.

14 The KT impact site identified:the location of the Chicxulub crateron the Yucatán Peninsula, Mexico,partially underwater, with thecoastline as it was at the end of theCretaceous also shown.

The crater was located in a piece ofbrilliant detective work by a young

Canadian graduate student, AlanHildebrand.19 He, along with others atthe time, had realized that the crater mustlie somewhere in the Caribbean area.There were two main lines of evidence:tumbled rocks produced by ancienttsunamis, and indicators of proximalityin the boundary beds.

First, along the shores of easternMexico, and in Texas and other southernstates of America, geologists hadidentified unusual, chaotic beds at theKT boundary. Mixed with boundaryclays and the iridium were tumbledblocks of the local limestones. Theblocks had been torn up and dumped bysome dramatic physical event, and it hadbeen suggested that the crisis might have

been a huge tidal wave, a tsunami. Thesedimentological evidence has beendisputed for many of these chaotic beds,but perhaps an impact somewhere out inthe proto-Caribbean sea had sent out avast tsunami that pounded the shores?

Second, Hildebrand and others hadnoticed that the KT boundary layerbecame thicker towards the centre of theCaribbean. So, whereas the boundarybed was about 1 centimetre thick in mostparts of the world, in Texas and Mexicoit was a metre or more thick in places.Geologists are alert to indicators ofproximality: the closer you approach avolcano, the thicker the lava beds; thecloser you are to the head of a submarinesediment slide, the thicker the layers of

sediment. Not only thickness, but alsograin size. In the thicker boundary layersaround the Caribbean, geologists foundbuild-ups of glass beads, and these werenot found further afield.

Hildebrand found some boreholerecords that had been recorded by theMexican oil company Pemex in the1960s. The oil geologists had identifieda large circular structure deep beneathChicxulub, which they thought might bean oil trap. They drilled on either side ofthe structure, and right through themiddle of it (Fig. 15). To the sides, theboreholes penetrated 3 kilometres ofearly Tertiary sediments and underlyingCretaceous limestones. But in the middleof the structure, the drill first

encountered late Tertiary sediments,then early Tertiary, and then noCretaceous. Instead of the expectedCretaceous limestones, they hit a brecciaand then a strange melt rock. From thedescriptions, Hildebrand guessed thatthe melt rock was suevite. He had hisevidence to identify the impact site.

The story seemed simple. Themeteorite had struck in the proto-Caribbean, evaporating the water andpunching kilometres down into theCretaceous limestones and salt beds.Great boulders and debris were hurledin the air and fell around the crater. Meltproducts – glass bombs and beads –flew through the air and spread over anarea at least 1000 kilometres across.

The finer material entered thestratosphere, and transported shockedquartz, ash and iridium around the entireglobe.

15 Cross-section of the Chicxulubcrater, showing the shape of thecrater from geophysical surveys.The suevite (melt rock) and brecciabelow the crater have beenrecovered from boreholes. Notehow the crater filled with sedimentafter the impact.

After the impact, sediments weredeposited over the crater and thesurrounding landscape. By mid-Tertiarytimes, some 40 million years after theimpact, there was still a noticeabledepression lying over the crater. LateTertiary sediments filled it up, butsediments of that age were stripped offthe surrounding landscape by normalerosion processes. Today, there is no

physical sign of the crater since theground is entirely level. In any case, halfthe crater lies offshore, beneath themodern Caribbean; the only way to studyit is by geophysics.

Geophysical traverses of theChicxulub crater show that it has atriple-ring structure. The inner ring,produced by the springback of theEarth’s crust within seconds of theimpact, is 80 kilometres across. A zone,from 100 to 130 kilometres across,represents the original crater edge, and azone of collapse into the crater itself.Finally, at a diameter of 195 kilometres,is an outer ring which probably marksthe extent of the force zone of the impact.It seems that the vast energy of the

impact created a crater 130 kilometresacross, but also pulled down a circle ofcrustal rocks 195 kilometres across.

Where are we today?

The KT event is not completelyresolved. One thing is clear, however:there was an impact centred onChicxulub in Mexico 65 million yearsago (radiometric dating confirmed theage of the melt rocks). This impact hadall the global effects predicted fromclose study of the Ries crater, andmodelled by Luis Alvarez and hiscolleagues in 1980. The impact seems tohave sent out vast tsunamis that had

devastating effects in the proto-Caribbean, and ash and iridium circledthe globe. The ash certainly blocked outthe sun, leading to global darkness andfreezing. Evidence for freezing comesfrom studies of plant stems and leavespreserved at the moment of impact: thecell walls have been burst by growingice needles in the sap. But did this causethe extinctions?

Some of the criticisms offered bypalaeontologists in 1980 are still validtoday. Many groups of organisms wereindeed in decline before the impact, andthese declines may relate todeteriorations in climate or changes insea level. It is important also to recallthat many plants and animals were

seemingly unaffected by the impact, sothe killing model has to take account ofthat. Ever more detailed studies of fossiloccurrences up to the KT boundary mayshed further light on what was going on.

In addition, there were also majorvolcanic eruptions occurring at the sametime. The Deccan Traps in India cover ahuge area in the north of the country, andthey erupted over about a million yearsof time spanning the KT boundary. Thesewere basalt eruptions, that producedhuge volumes of lava in a number ofbursts of activity. Associated with theerupting lava would have been vastvolumes of carbon dioxide, sulphurdioxide and other gases. At one time, agroup of geologists suggested that there

had not been an impact at all, and thatthe Deccan volcanics could explain it all– the global ash layer, the iridium andthe rest.

This position is clearly untenablenow, since volcanic eruptions cannotexplain the shocked quartz or the glassbeads with geochemical compositions oflimestone. But there is no doubt that theDeccan eruptions happened, and thatthey began at a time of climaticdeterioration and decline of many fossilgroups. Maybe – and this is not yetresolved – the eruptions did set theextinctions rolling, and the asteroidimpact finished them off?

What a changed scientific world inthe course of 20 years! In 1980, despite

the work of the craterologists and thesuggestion of a supernova explosion 65million years ago, most earth scientistswere still firmly in Charles Lyell’scamp. When I learnt my geology in the1970s, my professors did not evenmention impacts, craters or massextinctions.

N o w , my students hear aboutcatastrophes, asteroids, giant eruptions,death and destruction every week intheir lectures. But were all massextinctions like the KT event? How doesit fit into the grand scheme of the historyof life, and how does it compare withthe even bigger end-Permian massextinction?

6

DIVERSITY, EXTINCTIONAND MASS EXTINCTION

Charles Lyell’s famous dictum ‘thepresent is the key to the past’ isdrummed into all trainee geologistsduring their first lecture. This injunctionmakes sense. If a geologist means to tryto understand events and processes thathappened in the past, it is self-evidentthat he or she should compare them with

what we know of the world today. Thebest way to understand the productionsof ancient volcanoes is to examine asmany modern volcanoes as possible. Butthere are some events and processes thatare so unusual, or so huge, that theycannot be studied in the modern world.

The grand pattern of the evolution oflife over millions of years is too big andtoo long-term for any biologist tounderstand simply by looking at livingplants and animals. Mass extinctions,great impacts and world-scale volcaniceruptions happen only rarely – perhapson time-scales of tens or hundreds ofmillions of years – and again studies ofthe puny events that we witness fromtime to time can only give hints of what

the larger catastrophes were like.Indeed, for studies of modernbiodiversity and extinction, the past canbe an essential guide to the present.

Life today is hugely diverse, and thatdiversity has been achieved in fits andstarts. There have been times of rapidrises in diversity, for example whenanimals with skeletons invaded theoceans, when life moved on to land,when the first coral reefs and the firstforests developed, and so on. But therehave also been many times of extinction,times when global diversity fell back.Georges Cuvier actually got it rightwhen he wrote in 1812:1

Thus life on earth has often been disturbed by

terrible events: calamities which initiallyperhaps shook the entire crust of the earth to agreat depth, but which have since becomesteadily less deep and less general. Livingorganisms without number have been thevictims of these catastrophes.

As we have seen, it was remarks likethese that so infuriated Charles Lyell.But Cuvier, and the stratigraphers likeMurchison, knew that there were markeddisjunctions in the record of the rocks,and in the record of life – times when allthe typical fossils of one time unitseemed to disappear, to be replaced byanother assemblage. As Cuviersurmised, many of these do indeedindicate times of extinction. But we nowknow that there were extinctions and

extinctions.

The big five

Five extinction events during the past500 million years stand out from the rest.These particularly large extinctions,commonly called mass extinctions, havebeen termed the ‘big five’. The end-Permian mass extinction was the biggestof all, and it has been called the ‘motherof all mass extinctions’ by Doug Erwin,2

who has written a great deal about it.The four smaller mass extinctionsinclude the KT event, and much studiedas it is, the KT event was by no means

the largest. The other three massextinctions happened in the LateOrdovician, the Late Devonian and at theend of the Triassic, some 440, 370 and200 million years ago respectively. Inaddition there were dozens of smallerevents, such as the loss of largemammals at the end of the ice age,10,000 years ago.

In order to understand massextinctions, it is important tocomprehend the concepts of time anddiversity. These are the two axes of thegraph: geological time, measured inhundreds of millions of years, is the xaxis, and diversity, measured in numbersof species or families, is the y axis.Extinction happens all the time, and

there is an expected, or background rate.Then, from time to time, somethingextraordinary happens, and there is ahigher-than-normal rate of extinction.This might then produce an extinctionevent or, if it is big enough, a massextinction.

Some mass extinctions have beenexamined in great detail, but there aremany questions still to be resolved. Forexample, the definition of a massextinction is not entirely clear: just howbig and how fast does the event have tobe to qualify? How ecologically diverseshould the victims be? There are majorquestions about the quality of the fossilrecord, and just how well it representswhat really went on: how close can you

get to an event that happened millions ofyears ago? Can geologists see whathappened day by day, or should they stepback and determine what went on overlonger periods, and in a more fuzzy, butmore accurate, way? Are there anycommon factors to mass extinctions? Forexample, are some kinds of organismsmore or less likely to be victims? All ofthese issues are important in trying tounderstand mass extinctions in general,and as a grounding for a closeexamination of the end-Permian event.They also have important implicationsfor understanding the modernbiodiversity crisis.

Diversity and biodiversity

Biodiversity is one of the fastest-growing words in the English language,indeed in any language. The word wasintroduced in 1988 attached to a report3on the growing ecological crises onEarth, and since then it has entered thepublic consciousness. ‘Biodiversity’ canbe read nearly every day in thenewspaper, and it is used in thousands ofscientific papers and reports every year.‘Biodiversity’ means no more than theolder, simpler term ‘diversity’ asapplied to life.

Simpler in form, but not simpler inmeaning. There are a hundred and one

meanings of the term ‘diversity’.Diversity can be assessed as the numberof species in a single area, such as aforest or a lake, or it can be a larger-scale term, such as the number ofspecies in a country, a continent, eventhe world. Diversity can refer to just onegroup, such as the diversity of birds orthe diversity of grasses in a field or acounty, or it can refer to all groups, frommicrobes to mammals. Diversity can besimply a count of species, or a highertaxonomic grouping such as a genus or afamily. In some ecological contexts,diversity also includes an assessment ofabundance, the number of individuals ofa species in an area – an attempt torecord the richness of life. In other

cases, diversity includes a measure ofthe ecological range of organismsobserved, whether the fauna includeslarge and small animals, for example, ora more limited scope. Diversity isincreasingly given a genetic ormolecular connotation these days too.Biologists try to assess the amount ofunique genetic material in a speciescompared to others as a means ofdetermining where to expend mostconservation effort.

Palaeontologists generally use theterm diversity in the simplest way, tomean a count of numbers of species,sometimes termed species richness.Species can be assessed in a localsample of fossils that are well

preserved. However, for global studies,it is normal to climb up the taxonomichierarchy to genera or families. Theseare larger groups, each containingseveral, or numerous, species, and theyare simpler to compare from continent tocontinent. Their record is also morecomplete than that of species. Here aretwo issues, then, the taxonomichierarchy, and the issue of quality of thefossil record. We will look at importantissues of quality later in this chapter.

Naming the beasts

From the earliest times, thinking peoplehave wondered at the diversity of life,

and they have named the plants andanimals they see around them. Such aninstinct dates from the earliest times, asis shown by the writers of Genesis(chapter 2, verse 19):

And out the ground the Lord God formed everybeast of the field, and every fowl of the air; andbrought them unto Adam to see what he wouldcall them; and whatsoever Adam called everyliving creature, that was the name thereof. AndAdam gave names to all the cattle, and to thefowl of the air, and to every beast of the field.

Naming and classifying are basic humanactivities. Roman citizens of twomillennia ago could name and identifyhundreds of plants and animals.Uneducated medieval peasants inEurope could do the same. In their

tropical forests, natives of the Amazonand of New Guinea identify the plantsand animals around them with greatprecision. Indeed, when biologists fromthe West visit these remote regions, theyfind that their formalized schemes ofclassification, based on decades ofcollecting and scientific study in thelaboratories of Europe and NorthAmerica, are no better than those of thelocals.

Biologists around the world haveagreed to a formal system of naming andrecording new species, and the formalstarting dates for published names are1753 for plants and 1758 for animals.4Those dates mark the publication of twofundamental books, the attempts by the

Swedish biologist Carl von Linné(Carolus Linnaeus, in the Latin form, thelanguage he, and scholars from allcountries, still used then to communicatemuch of their work) to instil order intothe classification of plants and animals –h i s Species plantarum (1753) andSystema naturae per regna trianaturae, secundum classes, ordines,genera, species, cum characteribus,differentiis, synonymis, locis (1758). Inthese, Linnaeus presented outlinedescriptions of all the plants and animalshe knew, an incredible achievement forthe time. However, the key contributionhe made was to establish a standard wayof naming organisms.

Linnaeus characterized each species

in his classifications by a two-part Latinepithet, such as Homo sapiens forhumans and Canis canis for the domesticdog. The first term, written with anupper case initial letter, is the genusname, the second the species. Eachgenus contains at least one species,usually several, sometimes many. Thespecies epithet is indicated with a lowercase initial letter. The whole name isalways given in italics.

The system is international and isrecognized by scientists worldwide,whatever their native language oralphabet. You see the Latin names ofspecies, written in the roman alphabet,in the middle of scientific texts inChinese or Russian. This makes for

perfect ease of communication betweennationalities, at least in terms of theformal names of organisms. Priority isimportant, and once a species has beennamed, no one else can give it a differentname, although a species may be movedto a different genus as a result of furtherstudy. The full form of a species epithetincludes the name of the author and dateof first publication, for example, Homosapiens Linnaeus, 1758.

From one species to many

New species evolve constantly, and lifehas diversified astonishingly since itsorigin some 3500 million years ago. In

that vast and unimaginable amount oftime, diversity, by definition, has gonefrom one species to many millions,perhaps 20–100 million today. Onespecies originally? Charles Darwin wasone of the first to dare to speculate, in1859,5 that all life had evolved from asingle species, or at most only a smallnumber. Where that species came from,he did not care to commit himself onpaper, and we need not enter into thevarious current hypotheses about theorigin of life. Nevertheless, one speciesit was. What is the evidence?

All living species, from the simplestsingle-celled virus or slimemould tohuman beings and oak trees, share themolecule DNA, which codes genetic

information, together with all thecomplexities of the protein manufacturesystem in the cells. In addition, all livingthings today share two furthercharacters: similar cell membranes withmechanisms to control the passage ofchemical ions and molecules in and outof the cells; and an energy transfersystem involving the molecule adenosinetriphosphate (ATP) by which energy isstored and released from a phosphatebond. These complex features of all lifeprobably arose once only, and hencethey point to a single ultimate ancestor ofall life.

What about the fossils? Of course,DNA, membranes and ATP moleculescannot be detected in the rock record.

But all fossils known to date belong toliving groups. Even the bizarre fossilgroups which are entirely extinct, suchas dinosaurs, hard-shelled fishes,trilobites, rhyniophyte plants and theweird beasts of the ancient Precambrianand Cambrian sea floors, clearly fitlower down within the evolutionary treeof modern forms. That is not to say thatlife did not originate more than once inthe primaeval ooze 3500 million yearsago, but any other experiments in lifehave gone undetected, or they led tonothing.

Expansion

How did life diversify after the origin ofthe first species? Some key steps can beidentified (Fig. 16). Many of the eventshappened in the long span of thePrecambrian, but these are the world’smost ancient rocks and so the fossilrecord is patchy, and radiometric datingis difficult. Fossils of tiny microbes areknown from some amazing, but veryrare, deposits, where they were instantlyentombed in glassy cherts. But suchchance discoveries, while hugelyimportant in documenting what went on,can give no true picture of diversityworldwide.

It is only by the end of thePrecambrian, during the Vendian period,that fossils become more abundant, and

rather larger. Then, from the beginningof the Cambrian period, some 540million years ago, fossils are found allover the place. Numerous groups ofanimals in the sea seemed to acquiremineralized skeletons of one sort oranother – shells or carapaces made fromcalcite, shells or bones made fromphosphate, or hard protein-rich cuticles– at about the same time. Again, this isnot the place to consider why that mighthave happened, merely to record that itdid.

Through the last 540 million years,termed the Phanerozoic (‘abundantlife’), life in the sea expanded indiversity in pulses. At first, marineorganisms swam or crawled on, or near,

the sea bed. Then, free-swimming formsbecame established, which fed on therich plankton at the sea’s surface. Largerpredators evolved to feed on the smalleranimals. Next, some groups began toburrow into the sea floor, seekingorganic matter. Others built up hugeskeletal structures called reefs. Throughtime, burrowers burrowed deeper, andreefs grew taller. Fishes evolved, thengreat marine reptiles to feed on them,then sharks, whales and seals, even seabirds, came on the scene.

16 The expansion of life, showingsix major innovations. Before theVendian (V) there had been overthree billion years of the history oflife, which saw the origin of life, ofcomplex cells and of multicellularlife. In the past 600 million years,one can note the Cambrian (C)explosion (1); the diversification ofreefs (2); the move of green plantson to land in the Ordovician (O)and Silurian (S) (3); the origin ofland vertebrates in the Devonian(D) (4); the spread of forests in theCarboniferous (Crb) (5); andeventually, the origin of humans inthe late Cenozoic (Cen) (6). Otherabbreviations: P, Permian; Tr,Triassic; J, Jurassic; Cret,Cretaceous; Tert, Tertiary.

Life expanded on to land rather later.Perhaps some simple algae spread agreen film over the watersides inPrecambrian time. But life did not moveon to land in a serious way until theOrdovician and Silurian, some 450 to400 million years ago. First, smallplants grew up around the sides of lakesand rivers. They were accompanied byancestral spiders, mites, millipedes andscorpions. Then came larger plants,insects and amphibians, the first landvertebrates. Ecosystems on land becamemore complex, with the evolution oftrees and forests. Insects diversifiedhugely, and so did land plants and landvertebrates. Evolutionary radiations ofconifers and seed ferns, dinosaurs and

pterosaurs, were followed by floweringplants and deciduous trees, mammalsand birds.

These patterns of expansion in thesea and on land have been documentedmany times by palaeontologists.6 Thesimplest way to show the patterns is by aplot of the numbers of families throughtime. Time here is taken as the last 600million years (myr.) or so, the Vendianplus Phanerozoic span, when fossilshave been abundant. Such graphs show(Fig. 17) how diversification wasepisodic. There was no uniformexpansion in a steady way through time.Bursts of diversification followed majorinnovations, such as skeletons or theability to fly, or the evolution of a major

new ecosystem, such as reefs or forests.Diversification has also sufferednumerous setbacks, some small, somelarger. The larger drops in diversity, ofcourse, mark the mass extinction events.

Extinctions and massextinctions

The fossil record becomes hazier thefurther back in time one goes. Therewere almost certainly numerous majorsetbacks in the history of life in thePrecambrian, but the precision of dating,and the abundance of fossils, are tooinadequate to be sure. The major

extinction events (Fig. 17c), includingthe big five, will be reviewed here intime order, starting with the oldest.7

The Late Precambrian event is ill-defined in terms of timing, but such anevent clearly occurred about 560–550myr. ago, when early kinds of animalsdisappeared. These early animals,sometimes termed a ‘failed experiment’,were found first in the Ediacara Hills ofAustralia, but have since been identifiedfrom around the world. The Ediacarabeasts include worm-like animals,probable jellyfishes, strange frond-likeforms and others – the first animalfossils that can be seen with the nakedeye. But the animals of the Ediacara typedisappeared, and the way was cleared

for the dramatic radiation of shellyanimals at the beginning of the Cambrianperiod (and the beginning of thePhanerozoic Eon).

17 Diversification of life in the sea(a), on land (b) and in total (c)through the past 600 million years,the Vendian and the Phanerozoic.Extinction events, major,intermediate and minor areindicated in (c). Abbreviations: C,Carboniferous; Cen, Cenozoic;Cm, Cambrian; D, Devonian; J,Jurassic; K, Cretaceous; O,Ordovician; P, Permian; Pc,Precambrian; S, Silurian; T,Tertiary; Tr, Triassic; V, Vendian.As John Keats wrote: ‘life is but aday; A fragile dew-drop on itsperilous way’.

A series of extinction eventsoccurred during the Late Cambrian,perhaps as many as five, in the interval

from 520 to 505 myr. ago. There weremajor changes in the marine faunas inNorth America and other parts of theworld, with repeated extinctions oftrilobites. Trilobites were complexanimals that looked superficially likewoodlice, with multiple limbs and anexternal hard skeleton. Their distantliving relatives include crabs, insectsand spiders. Following the series of LateCambrian extinctions, animals in the seabecame much more diverse, and groupssuch as articulate brachiopods, corals,fishes, gastropods and cephalopodsdiversified dramatically. Thegastropods, including modern snails,slugs and whelks, and the cephalopods,including modern squid and octopus, are

both subgroups of molluscs, the mostdiverse group of shelled animals.

In the Late Ordovician, about 440myr. ago, further substantial turnoversoccurred among marine faunas. Thisextinction event is the first of the ‘bigfive’ mass extinctions. All reef-buildinganimals, as well as many families ofbrachiopods, echinoderms (sea urchins,sea lilies and starfishes), ostracods(microscopic crustaceans, distantlyrelated to crabs and shrimps) andtrilobites died out. These extinctions areassociated with evidence for majorclimatic changes. Tropical-type reefsand their rich faunas lived around theshores of North America and other landmasses that then lay around the Equator.

Southern continents had, however,drifted over the south pole, and a vastphase of glaciation began. The icespread north in all directions, coolingthe southern oceans, locking water intothe ice and lowering sea levels globally.Polar faunas moved towards the tropics,and warm-water faunas died out as thewhole tropical belt disappeared.

The second of the big five massextinctions occurred during the LateDevonian, and this appears to have beena succession of extinction pulses lastingfrom about 370 to 360 myr. ago. Theabundant free-swimming cephalopodswere decimated, as were theextraordinary armoured fishes of theDevonian. Substantial losses occurred

also among corals, brachiopods,crinoids (sea lilies), stromatoporoids(colonial sponge-like animal colonies),ostracods and trilobites. Causes couldbe a major cooling phase associatedwith anoxia (loss of oxygen) on the seabed, or massive impacts ofextraterrestrial objects. The LateDevonian extinctions have figured, onand off, in the heated debates aboutwhether meteorite impacts cause massextinctions or not.

The largest of all extinction events,t h e end-Permian or Permo-Triassic(PTr) event, 251 myr. ago, is the third ofthe big five. The dramatic changeover infaunas and floras at this time has longbeen recognized, and was used to mark

the boundary between the Palaeozoicand Mesozoic eras. During the end-Permian event, most of the dominantPalaeozoic groups in the seadisappeared, or were much reduced:corals, articulate brachiopods,bryozoans (or ‘moss animals’, generallysmall colonial animals), stalkedechinoderms, trilobites and ammonoids(cephalopod molluscs with coiledshells). There were also dramaticchanges on land, with widespreadextinctions among plants, insects,amphibians and reptiles, which led in allcases to dramatic long-term changes inthe dominant replacing forms. Thecauses will be explored later.

The Late Triassic events include the

fourth of the big five mass extinctions. Amarine mass extinction event at theTriassic-Jurassic boundary, 200 myr.ago, has long been recognized by theloss of most ammonoids, many familiesof brachiopods, bivalves, gastropodsand marine reptiles, as well as by thefinal demise of the conodonts, ashadowy group of primitive fishes thatlived throughout the Palaeozoic, butwhich are known almost exclusivelyonly from their hardened tooth plates. Anearlier event, near the beginning of theLate Triassic, 225 myr. ago, also hadeffects in the sea, with major turnoversamong reef faunas, ammonoids andechinoderms, but it was particularlyimportant on land. There were large-

scale changeovers in floras, and manyamphibian and reptile groupsdisappeared, to be followed by thedramatic rise of the dinosaurs andpterosaurs, the flying reptiles. At thistime, many modern groups arrived on thescene, such as turtles, crocodilians,lizard ancestors and mammals. Thecause of these events may have beenclimatic changes associated withcontinental drift. At this time, thesupercontinent Pangaea was beginning tobreak up, with the unzipping of theCentral Atlantic between North Americaand Africa. Although meteorite impacthas again been pointed to as a possiblesuspect, the evidence is weak.

Extinctions during the Jurassic and

Cretaceous periods were minor inextent. The Early Jurassic and end-Jurassic events involved losses ofbivalves (two-shelled molluscs, likeclams, mussels and oysters), gastropods,brachiopods and free-swimmingammonites as a result of major phases ofanoxia. Free-swimming animals wereunaffected, and the events areundetectable on land – they may bepartly artificial results of incompletedata recording. Events have beenpostulated also in the mid-Jurassic andin the Early Cretaceous, but they arehard to determine. The Cenomanian-Turonian extinction event, 94 myr. ago,associated with extinctions of somefloating planktonic organisms, as well as

the bony fishes and ichthyosaurs(dolphin-like marine reptiles) that fed onthem, may turn out to be an artifact of amajor gap in the rock record.

The Cretaceous-Tertiary (KT) massextinction, 65 myr. ago, is by far thebest-known of the big five events. Aswell as the dinosaurs, the flyingpterosaurs, the marine plesiosaurs andmosasaurs, the ammonites andbelemnites (both common cephalopodgroups in the Mesozoic), various majorreef-building groups of bivalves andmost planktonic foraminifera(microscopic abundant shelly plankton)disappeared. The postulated causes, aswe saw in the previous chapter, rangefrom long-term climatic change to

instantaneous wipeout following a majorextraterrestrial impact.

Extinctions since the KT event havebeen more modest in scope. TheEocene-Oligocene events, 34 myr. ago,are marked by extinctions amongplankton and open-water bony fishes inthe sea, and by a major turnover amongmammals in Europe and North America.Later Tertiary events are less well-defined, which is surprising in a waysince the rock and fossil recordgenerally improves towards the present.There was a dramatic extinction amongmammals in North America in the mid-Oligocene, and minor losses of planktonin the mid-Miocene, but neither eventwas large. Planktonic extinctions

occurred during the Pliocene, and thesemay be linked to disappearances ofbivalves and gastropods in tropical seas.

The latest extinction event, at the endof the Pleistocene, while dramatic inhuman terms, barely qualifies forinclusion. As the great ice sheetswithdrew from Europe and NorthAmerica, large mammals such asmammoths, mastodons, woolly rhinosand giant ground sloths died out. Someof the extinctions were related to majorclimatic changes, and others may havebeen exacerbated by human huntingactivity. The loss of large mammalspecies was, however, minor in globalterms, amounting to a total loss of lessthan 1% of species.

There have been many extinctionevents, then, in the history of the Earth,the big five mass extinctions, and manysmaller events. Do the mass extinctionsshare any characteristics?

Background extinction andmass extinction

Extinction is normal. Species do not lastforever. Indeed, the average lifespan ofa species is perhaps 5 myr., with a rangefrom 100,000 years to 15 myr.,depending on what you are, whether amicrobe or a flowering plant. Speciescome and go and, even though the

overall diversity of life still seems to beincreasing, there is a steady rate ofbackground, or normal, extinction. Thebackground rate of extinction may beonly 10 to 20% of species per millionyears – 10 or 20 species out of every100 disappear every million years,which translates to one or two speciesper 100 every 100,000 years, or 0.01–0.02 per 100 per 1000 years, or0.00001–0.00002 per 100 per year. Tinyrates when measured in human terms, butmore noticeable on geological time-scales of millions of years.

Background extinction, extinctionevents, and mass extinctions. There haveclearly been times when extinction rateshave gone up, times that stand out as

extinction events of some kind.Extinction events are usually restrictedin some way, perhaps the loss of all thelarge, cold-adapted mammals at the endof the Pleistocene, or the loss of life onsome Pacific islands as a result of ahuge volcanic eruption or othercatastrophe. A mass extinction is thelargest kind of extinction event. It isimportant to try to define the meanings ofthese terms, not least because manybiologists have claimed that we areliving through the sixth, human-induced,mass extinction today.

The big five mass extinctions sharemany features in common, but differ inothers. Three things happened in all themass extinctions of the past:

• many species became extinct,generally more than 40 to 50%

• the extinct forms span a broad rangeof ecologies, and they typically includemarine and non-marine forms, plants andanimals, microscopic and large forms

• the extinctions all happened within ashort time, and hence relate to a singlecause, or cluster of interlinked causes.

These points all seem clear, butpalaeontologists have struggled to bemore precise. Just how many speciesshould disappear, and how fast, for anevent to stand apart from backgroundextinctions as a mass extinction?Attempts have been made to find a more

quantitative definition of whichextinctions are truly mass extinctions,and which are more localized orecologically restricted events, but noneof these efforts has really beensatisfactory.

A statistical test?

In 1982, the palaeontologists DavidRaup and Jack Sepkoski of theUniversity of Chicago8 claimed that theyhad found such a test for massextinctions, and their idea was simple.They argued that if times of massextinctions were associated with

exceptionally high rates of extinction,then these should stand out clearly fromnormal background extinction rates.Raup and Sepkoski calculated a meanrate of disappearance of marine animalfamilies, per million years, for eachgeological stage (average duration, 5–6myr.). So, over a timespan of the last600 myr., they assessed about 100separate extinction rates.

Raup and Sepkoski believed thatthey could apply a simple statisticaltechnique to these measurements:regression analysis. This is a grand termfor fitting a line to a set of points on agraph. The ideal is that a straight linecan be fitted, and that most of the pointswill lie close to it. There was absolutely

no reason why Raup and Sepkoski couldhave fitted a line to their measurementsof extinction rates through time, but theysucceeded. Most of their 100 extinctiondata points fell very close to a straightline, and that straight line was decliningthrough time (Fig. 18). In other words,they argued, the likelihood of extinctionwas diminishing, which they attributedto ‘improvements’ in the ability oforganisms to avoid extinction. What ofthe points that did not fall close to theregression line? These turned out tocorrespond to the big five massextinctions: Late Ordovician, LateDevonian, end-Permian (PTr), LateTriassic and Cretaceous-Tertiary (KT).

Statistically, the case seemed clear.

In formal terms, the regression line fittedthe data with 95% confidence, meaningthat at least 95% of the points lay veryclose to the regression line. Theremaining 5% of points, five in this casesince the total number was 100, aretermed statistical outliers. They lie awayfrom the general trend, and must beexplained in some other way. Perhapsthey were wrongly measured or, theinterpretation generally accepted,perhaps they were truly different fromthe other points. If so, that would meanthat mass extinctions had been identifiedin a quantitative way, and they had beenset apart. The implications would beprofound if the observations werecorrect.

18 Total extinction rate, measuredas extinctions per million years, formarine invertebrate families. Mostpoints fall on either side of aregression line that declinesthrough the last 600 million yearsof the good-quality fossil record.Five sets of statistical outliers,lying above the 95% confidenceenvelope (enclosed by dashedlines), correspond to the five majormass extinctions.

The five mass extinctions had beenpicked out as something different fromnormal extinction, certainly much bigger,but perhaps more than that. Raup andSepkoski hinted clearly that massextinctions were not only quantitativelydifferent from normal extinction events

(i.e. bigger), but perhaps they were alsoqualitatively different. A whole differentclass of phenomena. If they were right,this would suggest that the normal rulesof evolution broke down during massextinctions, and that unique kinds ofcauses should be sought. Perhaps,indeed, all mass extinctions mightthemselves share characteristics witheach other, and not with other smallerextinction events. This important ideawas actively discussed, but ultimatelyrejected, by most palaeontologists in the1980s and 1990s.

The implications of Raup andSepkoski’s 1982 paper were profound,but these authors had to admit, whenchallenged by a statistician, that the

method was flawed. The technique ofregression analysis could not be appliedto their data.9 The idea was strong, butthe statistics were weak. None the less,their studies led Raup and Sepkoski tothink more about the fact that there hadbeen repeated mass extinctions throughtime: what if they followed a regularpattern?

Periodicity, and the nextasteroid to strike the Earth

Could all mass extinctions be explainedby a single cause? Some geologists hadalready suggested that the history of life,

including phases of diversification andphases of extinction, might be controlledby changes either in temperature (usuallycooling) or in sea level. The idea wassimply that the physical environment onthe Earth was sometimes attractive forlife, and at other times it was not.Geologists in particular found such ideaseasy to accept, that life was merely apart of the huge complex of the Earth’ssystems, an intimate weaving of thegeosphere and the biosphere. Biologiststend to find such ideas of externalcontrols on large-scale evolution harderto accept, since they are used to seeingchanges brought about by interactionsamong species, for example competitionor predation. The contrast is between

external control, the geologists’ view,and internal control, the biologists’view.

The revolution in KT studies broughtabout by the Alvarez paper of 1980clearly set palaeontologists andgeologists thinking. What if the history ofthe Earth, and in particular the history oflife, were controlled in certain ways byrepeated impacts of asteroids or comets?

The search for a common causegained great credence with the discoveryin 1984 of an apparently regular spacingbetween mass extinctions during the last250 myr. Raup and Sepkoski found aregular period of 26 myr. separatingpeaks of elevated extinction rates for therecord of marine animals (Fig. 19).

Initial responses to this proposal werepolarized. Many enthusiastic geologistsand astronomers accepted the idea, andits clear implication that a regularperiodicity in mass extinctions implied aregular astronomically controlledcausative mechanism. It is worth pausinghere for a moment.

19 Periodic mass extinctions? Thisis the famous diagram from Raupand Sepkoski (1984) that launcheda thousand astronomical studies. Ifthere is a regular 26-million-yearperiodicity, an astronomical causeseems essential, whether it betilting galaxies, a twin starNemesis or Planet X.

If Raup and Sepkoski were correct,their simple palaeontological studywould imply a whole new galaxy-scaledsystem of processes that controlled theentire history of the Earth and of life.Also, certain people thought, if we knowwhen the last crisis happened, we cantherefore work out when the nextmonster asteroid will hit. The Alvarez

proposal of 1980, that the Earth hadbeen hit by an asteroid 65 million yearsago, was nothing when compared to thegrandeur of the new proposal by Raupand Sepkoski.

Numerous astronomical models toexplain the periodicity of massextinctions were proposed. All thesemodels involved a regularly repeatingcycle which disturbed the Oört cometcloud and sent showers of cometshurtling through the solar system every26 myr. The Oört comet cloud is a shellof comets at the fringes of our solarsystem, lying beyond Neptune and Pluto.The disturbing factor might have beenthe eccentric orbit of a sister star of theSun, dubbed Nemesis (but not yet seen),

tilting of the galactic plane or the effectsof a mysterious Planet X which laybeyond Pluto on the edges of the SolarSystem.

The astronomers enthusiasticallypolished their telescope lenses andpointed them skywards in search ofNemesis, Planet X or tilting galaxyedges. Geological supporters ofperiodicity went out to find evidence formassive impacts at mass extinctionboundaries to match the physicalevidence that had already beenestablished for the KT event. But thecritics of periodicity argued that eachmass extinction was a one-off, and thatthere was no linking principle. The 26myr. cycle discovered by Raup and

Sepkoski was, they argued, a statisticalartifact or the result of limited dataanalysis.

So what is the current view ofperiodicity? I think that mostpalaeontologists and geologists have justquietly let it drop. Close analysis of thefossil data has failed to confirmperiodicity. Indeed, scrutiny of some ofthe extinction peaks in Fig. 19, such asthe three in the Jurassic, has suggestedthat these are largely artifacts of the datacollecting. Also, the searches forNemesis and Planet X have not beensuccessful; nor has the search forindicators of impact at the time of theother identified mass extinctions.Iridium, shocked quartz or craters have

been found for only two or three of theten postulated mass extinction peaks thatare elements of the periodic cycle, butthis evidence is feeble in the extremewhen compared to the manifold lines ofevidence for impact at the KT boundary.

Periodicity of mass extinctions wasan intriguing idea, but it has been almostconclusively rejected now. But that doesnot mean that one should not take anoverview of all the extinction events ofthe past in search of common factors.

Scaling and taxonomictargets

Extinction events of the past vary inmagnitude, and they may usefully besorted into major, intermediate andminor events, based on their magnitudes(Fig. 17c). The end-Permian massextinction is in a class on its own, sinceit is known that 60–65% of familiesdisappeared at that time, and this scalesup to a loss of 80–95% of species. Thefour intermediate mass extinctions areassociated with losses of 20–30% offamilies, and perhaps 50% of species.The minor extinction events experiencedperhaps 10% family loss and 20–30%species loss, but these cannot be calledmass extinctions.

What do these species loss figuresmean? Are we really sure? Well, in

assessing mass extinctions,palaeontologists generally rely on largersubsets of the classification of life. Itwould be marvellous if the fossil recordwere complete, and all species that everexisted were laid out to be seen andrecorded. There would then never beany doubt about what happened, andwhen, and even why. However, mostorganisms that ever lived do not becomefossils. Indeed, perhaps even mostspecies that ever existed do not becomefossils, and this thought certainly givespalaeontologists sleepless nights.

Fortunately, there is an importantscaling issue that saves thepalaeontologist’s sleep: the bigger thetaxonomic target, the more chance you

have of hitting it. As you climb thetaxonomic hierarchy, the chances ofpreservation increase. So, the chance offinding any particular fossil speciesmight be one in a hundred, or 1%. But ifthere are ten species in a genus, then thechance of finding that genus increases to10%. If there are ten genera in a family,the next traditional rank in the taxonomichierarchy, then the chance of finding atleast one representative of that family is100%. So, palaeontologists make theirinitial global studies at the level of thefamily. Once that survey is complete,they may attempt to drop to the genuslevel, or even to the species level,perhaps for local studies.

It is possible to use this scaling logic

in reverse to interpret the family-leveldata to the species level. In every case,it is clear that a higher proportion ofspecies than families are wiped out inany extinction event, simply becausefamilies contain many species, all ofwhich must die for the family to bedeemed extinct. Hence, the loss of afamily implies the loss of all itsconstituent species, but many familieswill survive even if most of theircontained species disappear. So, a lossof 60% of families in the end-Permianevent must mean a loss of a higherproportion of the species, and that hasbeen variously estimated as a loss at thelevel of 80–95%.

Seeing what you want to see

The first defining character of massextinctions is that many species shouldhave disappeared. In broad terms, thescale of extinctions ranges from lossesof 20–60% of families, and estimatedlosses of 50–95% of species overall. Indetail, however, these rates are notuniform across all groups: extinctionrates vary from 100% of species withinclades that disappear entirely to 0% ofspecies in other clades. The term ‘clade’has recently come into use more andmore, and it is a useful expression. Itrefers to a group of any size that had asingle ancestor, and includes all of itsdescendants. So, Homo sapiens is a

clade, as is the Family Canidae (dogsand relatives), the Class Aves (birds),the Kingdom Animalia, and so on.

Good-quality fossil records indicatea variety of patterns of extinction.Detailed collecting of planktonicmicrofossils based on centimetre-by-centimetre sampling up to, and across,crucial mass extinction boundariesoffers the best evidence of the patternsof mass extinctions. In detail, some ofthe patterns reveal rather suddenextinctions (Fig. 20). Study of the rocksuccession across many KT boundariessuggests that the sequences are ascomplete as could ever be achieved, andthat the total time intervals involved arein the range of 0.5–1 myr. The precision

of dating may be as fine as 50,000 yearsin these cases. This is amazing accuracyfor a geologist, but pretty hopeless for abiologist.

Does this superb field example (Fig.20) show a catastrophic or gradualpattern of extinction? The problemconcerns definitions: how sudden issudden? Palaeontologists with differentbiases could argue that Fig. 20 shows acatastrophic or a gradual dying. Agradualist would argue that there areseveral steps of extinction over 50,000years or more, and they are far toodrawn out to be the result of an instantevent. A catastrophist, on the other hand,would see instant extinction occurring inas short a time as 1–1000 years, and

would argue that the slightly steppedpattern is the result of incompletepreservation, inadequate collecting ormixing of the sediment by burrowingorganisms which churned up the mudbefore it hardened into rock.

20 A catastrophic pattern ofextinction. The recorded time-spans of different species offoraminifera in the classic KTsection at El Kef, Tunisia, based oncentimetre-by-centimetre samplingthrough 5 metres (about 1.5 millionyears) of sediments, show a speciesloss of 65% at the KT boundary.

What is to be done? Will detailedstudies of mass extinctions always leadto this kind of stand-off, wherecatastrophists and gradualists see whatthey want to see? Many groups oforganisms do indeed seem to have goneextinct geologically rapidly during massextinction events, perhaps in spans oftime less than 0.5 myr., but that is still

far too long an interval for properbiological interpretation since it couldencompass gradual or catastrophic kindsof causes. The dispute always comesdown to problems of sampling.

Sampling the death of thedinosaurs

Fossil sampling is a key issue. Even if apalaeontologist can prove that dinosaurfossils suddenly disappear from the rockrecord at a particular horizon, it cannotsimply be assumed that thedisappearance is the result of extinction.There might have been an environmental

change at that point, and the animalstrotted off elsewhere, or there mighthave been a substantial break indeposition, or depositional processesmight have changed in such a way thatbones are no longer buried andpreserved.

Dinosaur fossils, after all, arepreserved in continental sediments(sediments deposited on land, in riversor in lakes). Continental sediments arenowhere near as continuously depositedas are sediments in the deep ocean.Specimens are large and rare, so thatdetailed bed-by-bed sampling is fraughtwith difficulties. Is there any kind ofsampling regime that might allow apalaeontologist to resolve these

problems?Two separate teams have attempted

to resolve the question of the timing ofdinosaur extinction once and for all.This might seem unnecessary, since mostpeople accept that the meteorite impactat the KT boundary was a sudden event,as discussed in the last chapter.However, it would be wrong then toassume that the dinosaurs died outovernight as a result. That is a separateclaim that has to be demonstrated. So,both teams attempted large-scalecontrolled field sampling in Montana,and they amassed thousands ofspecimens from the last 5–10 myr. of theCretaceous rocks of the Hell CreekFormation.10 Needless to say, one team,

led by David Archibald of San DiegoState University, found evidence for along-term die-off, and the other team, ledby Peter Sheehan of Milwaukee PublicMuseum, found evidence for suddenextinction. Each sampling exercise hadinvolved teams of dozens of people,logging in one case an estimated 15,000person-hours of field prospecting, and inthe other case a total of 150,000identified specimens: how much moreintensive does the programme have to bein order to establish what reallyhappened?

Surely palaeontologists can at leastanswer that one simple question: did thedinosaurs die out instantly or over a longtime? There are two main problems,

however, which prevent a clearcutsolution: dating and selectivepreservation. First is the problem ofprecise dating. In some places,radiometric dates at, or close to, the KTboundary have been measured, but theyhave an inbuilt error of at least a fewhundred thousand years. And then thereis the problem of correlation, thematching of rock ages from place toplace. Can we really be sure that the lastrocks with dinosaurs in Montana, say,are the same age as the last rocks withdinosaurs in Wyoming? Or, moresignificantly, are the last dinosaurs inMontana the same age as the lastdinosaurs in France, Romania orMongolia?

Until these questions can beanswered unequivocally, it will beimpossible to say with certainty that thedinosaurs died out instantaneouslyeverywhere, or over the course of half amillion or a million years. Selectivepreservation is also a critical issue.

Taphonomy

Selective preservation of fossils is a factof life. The preservation of fossilsdepends on the biology and ecology ofthe organisms, the vicissitudes oftaphonomic processes (everything thathappens between death andfossilization) and the style of collecting.

These factors can be thought of as aseries of filters that an organism has topass through before it can be treated as aknown part of the fossil record. Indeed,such is the severity of the filters that it isa wonder that any fossils are preservedat all!

Biology and ecology first. Thebuilding blocks of an organismdetermine critically whether it can everbecome a fossil. If there is a skeleton ofsome kind – bones, a calcareous shell,woody tissue in a plant – then theorganism may well be fossilizable. Softtissues, such as guts and muscles andskin, may be preserved in unusuallygood chemical situations, but that is rare.So, jellyfish and worms do not have

good fossil records since they lackskeletons. Ecology is important too.Organisms that live on the sea bed aremore likely to be preserved than fliersor tree-dwellers as more sediment isdeposited on the sea bed. Fast-breedingorganisms with short life cycles aremore often preserved, since there aremore of them to find.

Taphonomic filters are many andvaried. A skeleton is not a guarantee ofpreservation. Bones and shells areusually ground up by physical processes,say transport in a river and being rolledamong boulders. Or scavengers andmicrobes may completely eat up allbody parts of a carcass. Once in thesediment, bones and shells can be

dissolved by weak acids in undergroundwaters. Even if the potential fossilsurvives all these onslaughts, the rocksin which it becomes entombed maysubsequently be crushed, heated, erodedor otherwise mauled by earthmovements.

Human factors also come into play.Scientists are not automata. They do not,robot-like, record every fact that is to berecorded, and brave every difficulty toobtain those facts. If you look atpalaeontologists’ field maps, their findsoccur near roads, near towns, nearfacilities. I don’t wish to suggest thatpalaeontologists are slackers – themajority are enormous enthusiasts, whowork unspeakable hours in the field.

None the less, there are lots of fossils,and not so many palaeontologists, so theknown fossil record represents only apart of the whole fossil record that islocked up in the rocks. Is the knownfossil record seriously biased? If itwere, then palaeontologists would haveto tread with caution.

A biased fossil record?

Fortunately, a number of studies nowsuggest that published accounts of thefossil record are not seriously biased.For example, Des Maxwell, once myresearch student and now at theUniversity of the Pacific, in Stockton,

California, and I looked at howpalaeontological knowledge hadchanged over the past 100 years.11 Weselected some key palaeontologytextbooks published in 1890, 1933,1945, 1966 and 1987, and drew upcharts of the diversification andextinction of tetrapods. We found that,although the total number of knownfamilies had doubled since 1890,obviously as a result of the huge amountof fossil collecting during the twentiethcentury, the overall patterns had notreally changed (Fig. 21). Had he wishedto, the palaeontologist in 1890 couldhave identified the same massextinctions and times of rapiddiversification that we could see in

1990.In a similar study, Jack Sepkoski

looked in detail at how hisinterpretations of the marine fossilrecord had changed in the 10 yearsbetween his two ‘Compendia of marineanimal families’, published in 1982 and1992. The number of families hadincreased by perhaps 10%, but theincrease was random through time. Allthat had changed was that the massextinctions had become sharper. This isstrong evidence that palaeontologistshave a good understanding of the fossilrecord. I would make the prediction thatanother 100 years of fossil collectingwill not really change things very much.I’d be amazed if a fossil human is found

in Mesozoic rocks, or a dinosaur in theSilurian.

21 The known fossil record is agood sample of the entire fossilrecord. Intense collecting efforts bypalaeontologists over the past 100years have doubled the numbers offamilies recorded, but the overallpatterns of diversification andextinction have not really changed.The palaeontologist in 1890 couldhave identified the same events ashis counterpart in 1987.

These were global-level studies.What about the smaller scale, the local-level study? Peter Ward of theUniversity of Washington, Seattle,showed how his records of ammoniteextinction at the KT boundary12 changedwith more field collecting. He chose as

his field area the famous KT boundarysection in marine limestones at Zumayain northern Spain. In two field seasons,in 1984 and 1985, Ward crawled allover the beach and collected everyammonite specimen he could find.Specimens were rare, but he identified16 species, and drew up a chart of theiroccurrence. Some of the species seemedto last through only a few metres ofsediment, while others spanned tens ofmetres of thickness, and hence,presumably, longer amounts of time.Based on his 1984–85 field collecting,Ward thought he had found a gradualpattern of ammonite extinction. Allspecies were gone well below the KTboundary, and they had begun

disappearing many metres below.Ward revisited the Zumaya section

in 1987 and 1988, and spent longer,searching more intensively. He managedto find two new species and he extendedthe ranges of all the ammonites he hadfound by 1986. One species nowreached the KT boundary. In otherwords, by collecting harder, he hadfilled some gaps. Further collecting,over the whole region, from 1988 to1990, filled more gaps, and now 10 outof a total of 31 ammonite speciesreached the KT boundary. By intenseeffort, Ward had converted a gradualextinction into a catastrophic one. Wasthe extinction catastrophic? Probably.And here was a case where human effort

was important in establishing the truepicture.

The dilemma of the fossilrecord

Two apparently opposed viewpointsseem to be emerging. On the one hand, itis obvious that the fossil record is notcomplete, that many organisms that oncelived are not preserved. And it must beclear that the fossil record becomesworse and worse the further back in timeone goes. On the other hand, commonsense says that palaeontologists knowthe fossil record well. The recent studies

by Maxwell and by Sepkoski seem toprove that. The point is, of course, thatthe fossil record – all the fossils we canfind in the rocks – is merely a sample ofthe diversity of the life of the past. Sowe know the fossil record well, butperhaps the fossil record does not reallytell us the true history of life? That’s adisturbing thought – for a palaeontologistat least. It might suggest thatpalaeontologists can know the fossilrecord, but they cannot say much aboutthe history of life.

This thought has worried me forsome time. I am a palaeontologist, and Iwould like to think that I can say someuseful things about the history of life. Inparticular, I would like to think that the

fossil record can tell us about majordiversifications and major extinctionevents of the past. How can the problembe resolved? How can anyone comparethe known fossil record with the truehistory of life, unless one has a timemachine, or can read the mind of God?

The solution was found in the early1990s. Palaeontologists are fortunate inhaving more than one source ofinformation about the history of life.There are of course the fossils in therocks, but there is also an entirelyindependent source of data fromphylogeny. Phylogeny is the shape of theevolution of life, the evolutionary treesproduced by taxonomists when they tryto bring order into the diversity of

organisms.Until the 1980s, most taxonomists

used some principles of stratigraphy andfossil dates in making up theirphylogenies. They would arrange thefossils in order, and essentially join thedots. Older fossils were ancestors ofyounger fossils, and so on up to thepresent day. This method works wellwhen the fossil record of a group is verygood, when all the ancestors anddescendants are preserved. But that isoften not the case. Usually the fossilfinds are patchy, and they show only theoutlines of the pattern of evolution, notthe detail. Alternative methods wereneeded.

Drawing trees

Two new methods for drawing upevolutionary trees were developed in the1980s and 1990s: cladistics andmolecular phylogeny. Both had had theirorigins in the 1950s and 1960s.Cladistics grew out of the need toformalize taxonomic methods in someway. Willi Hennig, a Germanentomologist, realized that taxonomistsat the time were using all kinds ofcharacters to make their phylogenies andclassifications. They might study theshapes of skeletons and teeth, wings andspines, colours and patterns. This wasfine, but there was no strict test of whichcharacters might be helpful or not. In

trying to draw up a phylogeny of humansand apes, it is no use to note that theyhave hair and legs. All mammals havehair and legs. It is probably more usefulto look for finer features of the skeletonthat are shared by two or more speciesof humans and apes, and then to considerwhether any of these characters indicatesa unique acquisition in evolution.

The methods of cladistics,13

proposed by Hennig in the 1950s and1960s, provided a technique fordiscovering evolutionarily significantcharacters – features that indicated asingle acquisition event, a branchingpoint. In the human-ape case, significantcharacters include the relative brainsize, the broad chest, the lack of a tail,

the shape of the molar teeth: all of theseare shared by apes and humans, but notby monkeys. By trial and error, nowusually carried out by computer, allpossible patterns of relationship areassessed, until the one which bestexplains the data is arrived at. Thepattern is shown as a branching diagramthat demonstrates just how humans,chimps, gorillas and gibbons are related.And there has been no reference tofossils or stratigraphy. Fossils can beincluded, but their geological age doesnot affect the shape of the phylogeny.

The methods of molecular phylogenyreconstruction stemmed from the greatdiscoveries of molecular biology also inthe 1950s and 1960s. For the first time,

biologists could determine the precisestructure of biological molecules. Theydiscovered that organisms sharebiological molecules, but that theirstructure can vary subtly from species tospecies. At first, studies focused onproteins. The molecular biologistsdiscovered that, for example, thehaemoglobin of chimps was identical tothat of humans, and this confirmed a veryclose relationship (perhaps too close forsome). But human haemoglobin wasslightly different from that of a monkey,more different from that of a cow, andvery different from that of a shark. Theamount of difference in the proteins wasprobably proportional to the time sinceany pair of organisms shared a common

ancestor. This is the principle of theMolecular Clock: proteins change at afairly constant rate, and can thus give atime-scale of change and a pattern ofresemblances. In other words, thechemical differences could be calibratedagainst a time-scale, and phylogenetictrees could be drawn up.

Since 1990, attention has switchedfrom proteins to the nucleic acids, DNAand RNA, the key components of thegenetic code.14 It is possible to automatethe sequencing of DNA and RNA, andlong chains, often containing manygenes, can be compared among species.These sequencing methods lie behind theenormous data-handling efforts of theHuman Genome Project, the first fruits

of which were published early in 2001.There are now hundreds of molecularbiologists around the world whospecialize in turning sequenceinformation into phylogenetic trees.These phylogenies, like the cladisticones, are obviously entirely independentof stratigraphy. This is the key.

The test: confronting agesand clades

The first comparisons of informationfrom the independent sources ofstratigraphy (‘ages’) and phylogeny(‘clades’) were made in 1992 by Mark

Norell and Mike Novacek, both from theAmerican Museum of Natural History inNew York. 15 At the outset there is noclaim that either the phylogeny or thestratigraphy is correct: both are beingcompared to see whether they agree ornot. If they do not agree, then eitherphylogeny or stratigraphy could bewrong (there is, after all, only one truehistory of life). If they agree, then theymust be telling the true story. Couldn’tthey equally both be agreeing on thewrong story? That is not likely, sinceany biases that affect the quality of thefossil record (soft parts, poorpreservation, loss of specimens) cannotaffect the phylogenies, many of whichare based entirely on living organisms in

any case. What came of these studies?The 1992 comparisons showed good

agreement between phylogeny andstratigraphy in 75% of cases.Palaeontologists breathed a sigh ofrelief. The fossil record was telling agood part of the true history of life, andthe missing bits, the soft-bodiedorganisms, the beasts that lived in placeswhere fossils weren’t formed, did notoverwhelm the known parts of the fossilrecord. Various groups ofpalaeontologists have since pursued theidea, and the largest study, which Icarried out in Bristol with MatthewWills and Rebecca Hitchin,16 containedcomparisons of 1000 publishedphylogenies, cladistic and molecular.

The groups ranged from basal plants tobirds, trilobites to turtles, fungi to fishes.Many showed no real congruence, butthe majority did agree. Moreimportantly, and the point of the study,there was apparently no change in thequality of the fossil record through time(Fig. 22).

How can that be? Fossils fromCambrian rocks surely cannot be as wellpreserved as those from the Cretaceousor the Miocene. The older ones musthave been crushed, mangled, heated,covered or eroded much more than morerecent ones. Both points of view arecorrect – the damage to older forms andour results – and the resolution is interms of scope or focus, a point made

above. At the level of individualspecimens or species from a locality,there will obviously be a much poorercollection of Cambrian fossils thanCretaceous fossils. But at the higherlevels of genera and families, and overlonger time-spans, you are likely to findequivalent proportions of the truediversity from Cambrian and Cretaceousalike. True, you may have only 5specimens representing one of theCambrian families, and 10,000 for anequivalent Cretaceous family. But that’ssufficient.

These results are highlycontroversial. Have we really found away to compare the fossil record withthe true history of life? And that without

a time machine? We think so. If theresults are accepted, then we canconclude that the fossil record may notbe good, it may be full of holes, but it isevidently adequate to tell the true storyof the history of life.

22 Evidence that the quality of thefossil record has not changedthrough time. Two measures areused to compare the sequences offossils and the shapes ofphylogenies, SCI (stratigraphicconsistency index) and GER (gapexcess ratio). In both cases, lowvalues indicate poor matching, andhigh values good matching. Thisstudy is based on a sample of 1000molecular phylogenies andmorphological cladograms, andboth SCI and GER remainessentially constant through time.

Selectivity of massextinctions

If the fossil record is adequate, then wecan look again at mass extinctions forshared patterns, for evidence that mightbe useful in understanding how theyhappened, and, perhaps, to contribute tothe debate about the modern extinctions.It is hard to establish the timing of massextinctions, as we saw earlier, but itshould be possible to look at somebiological aspects. There are twoobvious questions: are mass extinctionsselective, and what happens after theextinction?

Selectivity first. The second definingcharacter of mass extinctions was thatthey should be ecologically catholic, thatthere should be no selectivity. This is asomewhat counter-intuitive proposition,

since most biologists would predict thatsome kinds of plants and animals aremuch more likely to become extinct thanothers. Surely elephants or pandas, orrare, isolated tropical trees are teeteringon the brink of extinction at all times?The commonest question I am askedabout dinosaurs is, ‘why did they goextinct?’ It is natural to try to findreasons. Surely extinction means failure,and the dinosaurs must have beenlacking some key faculty?

Numerous studies bypalaeontologists, however, have turnedup relatively little evidence forselectivity during mass extinctions. TheKT event certainly killed the dinosaursand some other large reptiles, but much

smaller animals died too, marsupialmammals, birds, some groups ofcrocodiles – so large body size was notthe sole weakness. Also, a large numberof microscopic planktonic speciesbecame extinct. It is not clear that nichebreadth (the range of habits and diets ofa species) was a strong factor either,since whole clades containinggeneralists and specialists disappearedat the same time. There is also noapparent bias towards extinction of topcarnivores.

The only evidence of selectivityduring mass extinctions has been againstspecies with limited geographic ranges.David Jablonski and David Raup of theUniversity of Chicago17 surveyed all the

bivalve and gastropod species andgenera of the latest Cretaceous andearliest Tertiary in North America andin Europe, and they found that generawhich were geographically restrictedwere selectively killed off, whencompared to those with wider speciesdistributions. So, if you want to survivemass extinctions, make sure you are aspecies within a geographicallywidespread genus. Body size andecological adaptations seem to beunimportant.

Precisely where you live might alsobe thought to be significant in selectivityduring mass extinction events. Forexample, it has long been suspected thattropical species are more extinction-

prone than are those with more polardistributions. This idea was based on theobservation that some mass extinctionevents are associated with an episode ofcooling. During such a cooling phase,temperate-belt forms could migratetowards the tropics, tracking their idealtemperature regimes, and they couldperhaps survive. But the tropical specieshave nowhere to go, as globaltemperatures plummet, and they aresqueezed out. Another recent study byDavid Raup and David Jablonski,18

however, has shown no evidence forlatitudinal differences in extinctionintensities of bivalves during the KTevent.

Recovery

What happens after a mass extinction?Clearly the survivors are a small subsetof the full pre-extinction diversity. Ifthere is very little evidence ofselectivity, then the missing species willcome from all over. There will be holesin the food webs on land and in the sea.In one place a key herbivore mightdisappear, in another a top carnivore.The survivors will evolve eventually toplug the gaps, or at least the wholeecosystem will eventually evolve backto full complexity.

After mass extinctions, the recoverytime, the time for diversity andecosystem complexity to rebuild, is

proportional to the magnitude of theevent. Biotic diversity took 10 myr. torecover after major extinction eventssuch as the Late Devonian, the LateTriassic and the KT. Recovery time afterthe massive end-Permian event wasmuch longer: it took some 100 myr. forthe diversity of families of plants andanimals around the world to recover topre-extinction levels.

It is possible to examine therecovery phase after mass extinctions inmore detail. One of the most-studiedexamples is the replacement ofterrestrial tetrapod faunas after the KTevent. With the dinosaurs and other landanimals gone, vertebrate communitieswere much impoverished. The placental

mammals, which had diversified a littleduring Late Cretaceous times, and whichranged in size up to that of a cat,underwent a dramatic radiation. Withinthe 10 myr. of the Palaeocene and EarlyEocene, 20 major clades evolved, andthese included the ancestors of allmodern orders, ranging from bats tohorses, and rodents to whales. Duringthis initial period, overall ordinaldiversity was much greater than it isnow: it seems that during the ecologicalrebound from a mass extinction,surviving clades may radiate rapidly,and many body forms and ecologicaltypes arise. Half of the dominantplacental groups of the Palaeocenebecame extinct soon after, during a

phase of filling of ecospace andcompetition, until a more stablecommunity pattern became established10 myr. after the mass extinction.

Lessons from the past

Much of the material in this chapter israther new research, certainly from thepast twenty years. The questions are big,and the solutions in many cases are stilltentative. I often think that the more westudy the fossil record, the less wereally know. But serious study of massextinctions is a very new science. It neednot have been – after all, John Phillipshad already documented some of the key

aspects over 150 years ago (Chapter 2 –‘The end-Permian mass extinctionrecognized’) – but the fear of beinglabelled a catastrophist ruled extinctionsand mass extinctions out of bounds forall reasonable scientists until 1970 or1980. So what do we know now?

• There have been several massextinctions in the past, at least five.

• Mass extinctions are characterized bya loss of 20–65% of families and 50–95% of species.

• There is some evidence that massextinctions may be set apart from normalextinctions as a separate class ofphenomena, but the evidence is limited.

• There is little evidence for selectivityduring mass extinctions, whether by size,diet or habits, though geographicallywidespread groups seem to be insulatedto some extent against the effects of massextinctions.

• Durations of mass extinctions seem torange from virtually instantaneous to 10myr. for some multiple events.

• Life has always recovered after massextinctions, but it took some 10 myr.after most of the mass extinctions forglobal levels of biodiversity to reachtheir pre-extinction levels, and as muchas 100 myr. after the huge end-Permianevent.

Many of these ideas about massextinction had been established by 1990,and yet at that time almost nothing wasknown about the biggest of all massextinctions, the end-Permian event.Considerable advances have been madesince then in teasing out just whathappened 251 million years ago. Thebest fossil record exists for life in thesea, so we will first explore just whathappened in the marine realm during thebiggest crisis for life of all time.

7

HOMING IN ON THEEVENT

In 1990, the American geologist CurtTeichert,1 in reviewing the end-Permianmass extinction event, wrote:

The way in which many Palaeozoic life formsdisappeared towards the end of the PermianPeriod brings to mind Joseph Haydn’s FarewellSymphony where, during the last movement,

one musician after the other takes hisinstrument and leaves the stage until, at the end,none is left.

This was the common view in 1990, thatspecies had died out piecemeal throughthe Mid and Late Permian, a span ofsome 10 million years. Extinction rateswere higher than normal, but perhaps anobserver at the time would not havedetected that they were actually in themiddle of the biggest crisis life had everfaced. The sum total of thedisappearance of species one-by-one iscomplete annihilation.

But is this the correct picture? Wasthe mass extinction at the end of thePermian so different from thegeologically instantaneous KT event? It

might seem unusual if this were the case.After all, the Late Permian event was somuch more serious in scale than the KT.Perhaps Curt Teichert, like manypalaeontologists at the time, was merelybeing cautious, and did not wish to pointto a catastrophic model?

In fact, this was largely not the case.The problem arose through difficulties indating the rocks immediately below andabove the Permo-Triassic boundary.Indeed it was not exactly clear where theboundary should be placed. Refinementsof dating can be traced through theestimates made by Doug Erwin of theSmithsonian Institution in Washington, aleading expert on the fate of marinemolluscs through the crisis. In his 1993

book,2 he estimated that the end-Permianevent lasted from three to eight millionyears. Only five years later, he and co-workers suggested a span of less thanone million years,3 and by 2000 thetiming of the extinction was less than500,000 years, and perhaps eveninstantaneous.4

The shift of opinion from gradualismto catastrophism in seven years was notmerely the replacement of old-fashionedwhims and prejudices. By 1990, workon the KT boundary had advancedenormously, and the likelihood of amajor impact and a rapid extinction atthat time were widely accepted. In fact,the tightening up of the end-Permianevent, from 10 million years to a

geological instant was mainly the resultof ever-more refined study of some ofthe geological sections that span thePermo-Triassic boundary.

But surely, by 1990, the Permo-Triassic boundary was well understood?After all, Roderick Murchison hadnamed the Permian in 1841, andgeologists had had 160 years in which topin down exactly where the boundarywith the Triassic lay. Surprisingly, theyhad failed to do so, and wranglingcontinued until 2000, when the ‘goldenspike’ was finally driven in, as we shallsee, at a section at Meishan in southChina.

Wrangling for a hundredyears

The story of the argument over definingthe Permian is pretty arcane, and mayseem pointless, but it is worthpersevering in order to appreciate thatthere is a method at work, thatinternational agreement is essential andthat in this case the turbulent politicalhistory of Russia hindered thisagreement. Clearly, it is essential tomake sure everyone is talking about thesame thing before one can begin to focuson a particular event, such as a massextinction, that might have lasted for justa geological instant.

Murchison had initially includedonly the middle and upper parts of whatwe now call the Permian in hisdefinition in 1841 – he in fact left out allthe lower part of the Permian, calling itgenerally Permo-Carboniferous, since hewas uncertain where to draw the line.So, Murchison’s5 Permian comprisedwhat is now assigned to the Kungurian,Ufimian, Kazanian and Tatarian stages inthe Russian system, just over half whatis now included (Fig. 23).

Both the top and bottom boundariesof the Permian had continued to bedisputed long after Murchison’s time. In1874, the Russian geologist A.Karpinskiy included the underlyingArtinskian and Sakmarian units in the

Permian.6 He showed that Murchisonhad identified these units wrongly whenhe trekked down the west side of theUrals. Murchison thought the marinesandstones and limestones around thetown of Artinsk were similar to theEnglish Millstone Grit, a familiar rockunit he had seen in northern England, anddefinitively part of the UpperCarboniferous. Karpinskiy’s studies ofthe fossils, however, proved that theArtinskian rocks had more in commonwith the overlying Kungurian, and wereclearly younger than Late Carboniferous.So they were assigned to the Permian.

23 Time-scale of the Permian,showing the internationaldivisions. The Lower Permian isbased on the Russian succession,the Middle Permian on the NorthAmerican (two Chinese equivalentsare noted), and the Upper Permianon the Chinese (two Russianequivalents are noted).

The true base of the Permian becameclearer only after more detailed studiesby Russian geologists andpalaeontologists around one hundredyears after Murchison’s work in Russia.V. E. Ruzhentsev restudied Karpinskiy’sSakmarian rocks and he realized thatthey fell naturally into two units, eachdistinguished by its own suites offossils. So he divided the Sakmarian intotwo, naming a lower Asselian stage in1950,7 and retaining the name Sakmarianonly for the upper portion.

During the nineteenth century, NorthAmerican geologists evolved their ownsystem for dividing their Permian rocks,and last century, the Chinese did just thesame. So now there are three

independent schemes available (Fig.23), and evidently this does not help inidentifying particular time lines on aworldwide scale. How is one to chooseamong these schemes?

The 1937 meeting inLeningrad

Why should Russia, North America andChina have their own systems of namingrock units? Obviously, nationalism is atplay here. Geologists desperately wantto have their national names usedworldwide – just as Murchison did inthe 1840s, when he had argued that it

was clearly right and proper that Britishsystems should be extended worldwidefor the universal benefit andimprovement of all foreigners.

There is also a geological reasonwhy it is difficult to come tointernational agreement on matching rockunits. Many different kinds of rocks arebeing deposited at any instant in time.For example, right now desert sands arebeing deposited in the Sahara, coral-richlimestones in the Caribbean, river sandsand muds in the Mississippi, soils andlake sediments in northern India, andglacial sands in northern Canada. Howwill the future geologist be able to tellthat all these sedimentary rocks are ofthe same age?

The problem is the same for Permianstratigraphers. The Lower Permian rocksof the Ural Mountains in Russia werelaid down mainly in the sea, and theycontain typical marine fossils; at thesame time, continental red beds wereaccumulating in Texas and New Mexico,complete with their faunas ofamphibians and reptiles. Inevitably, thefirst geologists who arrive in a newdistrict have to choose the marker bedsand fossils that are available. It is onlynow, some 160 years after Murchison’sfirst work in Russia, that the correlationsbetween Russia and China and NorthAmerica can be achieved with accuracy.In this case, politics played a part indelaying the necessary comparative

work.The key task of the stratigrapher is to

define the base of his or her rock units.(As mentioned above, unit tops areignored, since they are automaticallydefined by the base of the next overlyingunit – there might be even more disputesif tops were also defined, and did notsomehow correspond with thesuperincumbent unit base.) In naming thePermian System in 1841, Murchisonshould have defined the base of the newrock succession. This he believed he haddone in including the red beds and saltdeposits between Oka and Kazan, andaround Perm, on the western flanks ofthe Urals. We now know, however, thatthe huge thicknesses of fossiliferous

marine limestones and sandstones lyingbelow these units, and above definitiveCarboniferous rocks, are also Permian;he had omitted over 3 kilometres of rockaccumulation, corresponding to perhaps25 million years!

This omission was not merely anational affair, of interest solely toRussian geologists. It affected opinionaround the world. Americanstratigraphers were also puzzled, sincethey had to compare the fossils of theirextensive Permo-Carboniferous rocksuccessions with what they could findout about the Russian rocks. Until theRussians decided where to fix theboundary, the Americans could notmove.

It was only in 1937 that scholarsfrom around the world were able toexamine the Russian Permian seriously.This was at the height of Stalin’s rule inRussia, but the Seventeenth InternationalGeological Congress met in Leningrad(St Petersburg) that year, and a longfield trip was organized so that overseasgeologists could see the whole of theclassic Russian Permian.8 This was aunique experience for the American andEuropean visitors, familiar with theirown Permian successions, but desperateto examine the rocks that had givenMurchison the evidence to name thesystem. Equally, for the Russiangeologists, unable to leave their country,this was a chance at last to discuss their

work with scholars from other nations.The trip clearly made a huge

impression on Carl O. Dunbar,Professor of Paleontology andStratigraphy at Yale University, andChairman of the Permian Subcommitteeof the National Research Council’scommittee on stratigraphy. He wrote9 in1940 of ‘the never-to-be-forgotten daysof the Permian excursion’ in offering hisdetailed overview of what he had seen,really the first detailed account inEnglish since Murchison’s work onehundred years before. But he noted manyareas of disagreement. He reported howthe Russian geologists still disputedwhich rocks were to be correlated withthe European and North American

Carboniferous, and which should becalled Permian. What we now callAsselian he, and they, still assigned tothe Carboniferous.

The golden spike

Dunbar’s subcommittee was one of manythat had been set up to try to pin downinternational standards of definition ofgeological units. These committeespassed to an international level whenthey became part of the InternationalUnion of Geological Sciences in 1961,and of the International GeologicalCorrelation Programme in 1972, underthe umbrella of UNESCO. The role of

these committees, with input frominterested geologists of all nations, is tohammer in ‘golden spikes’ at the agreedboundaries of geological units. ‘Goldenspike’ is a figurative term; the boundaryis fixed precisely, once and for all, in aparticular location that can be visitedand studied – a spike made from gold isnot actually hammered into the rock.

In Murchison’s day, it was enough torefer in general terms to the rocks onewished to include in a newly named unit.By 1940, geologists were seekingsomething more precise, and theyrealized that the task would be difficult,since they had to survey the whole worldand choose the best available rocksection. To do this, they had to master

the intricacies of many sequences ofrocks, and the local nomenclatures fromdifferent lands. They had to attempt tocorrelate, or match, exact rock horizonsfrom place to place, from Russia toArgentina to China. They then had tochoose the geologically mostappropriate section to be designated asthe global stratotype, that is, the singlereference point that all geologists woulduse for ever more. Well, that’s the theoryof how it should work.

By agreement, golden spikes aredriven into boundaries in marinesuccessions, since fossils are moreabundant and sedimentation is usuallymore regular in shallow seas than onland. The stratotype section must be

thick and continuous, and without anyobvious major gaps. Fossils should beabundant and, ideally, of different kinds,such as ammonoids (coiled molluscshells) and conodonts (microscopictooth-like fossils that are usedextensively for dating marine rocks atthat time). Abundant fossils allow thepalaeontologists to fix the golden spikeimmediately at the first appearance of adistinctive fossil. Then, palaeontologistsaround the world can seek the firstappearance of this fossil species in theirlocal successions, and thereby make thecorrelation with some confidence. Ifammonoids are absent, then perhaps theywill be able to find the equivalentconodont that marks the same age of

rock.With the work by the Russian

stratigraphers, and by western geologistslike Dunbar, it might seem that theCarboniferous-Permian boundary couldhave been settled. But, after the SecondWorld War, contacts between Russianand western geologists became strained,and very few serious exchange visitswere possible. Many westerners tried toarrange further fieldwork in the 1950sand 1960s, but they rarely got furtherthan Moscow, where they would bepolitely taken on tours of the Kremlinand Red Square, but desperate requeststo visit the countryside and see somerocks were blandly refused, orpaperwork was misplaced.

The situation improved after 1990and, following many further meetingsand debates, the golden spike for thebase of the Permian was finally agreedin 2000, and it was placed in a streamsection on Aidaralash Creek, in theAktöbe (formerly Aktyubinsk) region ofnorthern Kazakhstan, lying some waysouth of the Ural Mountains. It is markedby the first occurrence of the conodontStreptognathodus isolatus. So, the baseof the Permian was finally agreed. Butwhat of the top, of much greater interestfor understanding the end-Permian massextinction event?

Seeking the base of the

Triassic

It became clear early on that the Russianrock successions would not beappropriate for fixing the base of theTriassic (and hence the top of thePermian). The uppermost units of theTatarian stage in the Urals, and betweenMoscow and Kazan, were largely redbeds, sandstones and mudstones laiddown in ancient rivers and as soilsbetween the river channels. Fossils werepresent, but these were mainly the bonesof ancient amphibians and reptiles, withassociated plant fossils and small pond-living creatures. Murchison may havebeen impressed by his finds of reptilebones (see Chapter 1), but modern

stratigraphers were not. The search wason for a marine succession that spannedthe Permo-Triassic boundary.

Much of western Europe was ruledout. The Permo-Triassic rocks ofBritain, France and Germany, forexample, were well enough known, butthere had been a major switch indeposition from the marine Zechstein tocontinental red beds in most areas.Indeed, there was a strong suspicion thatquite a span of time might also bemissing just at the boundary. In southernEurope, on the other hand, there weresome successions that spanned thePermo-Triassic boundary entirely inmarine rocks. Could these provide thekey?

In northern Italy, for example, in theDolomites, part of the southern Alps, therock successions show that sea waterswere generally deepening.10 The latestP e r m i a n Bellerophon Formationconsists of dolomites – alteredlimestones – that were deposited inlagoonal, subtidal conditions. Abovethese are oolitic limestones of theTesero Oolite Horizon. Oolites (theword literally means ‘egg stones’) aresmall limestone balls, made up from thinshells of limestone that precipitatedaround a sand grain or shell fragment,and they are formed generally in warm-water lagoons. The overlying MazzinMember consists of thinly bedded darkgrey limestones containing abundant

pyrite, evidence of further deepening ofthe water. The Permo-Triassic boundarylies within the Tesero Oolite Horizon.Below it are abundant latest Permianfossils; above it, these fossils haveapparently all disappeared.

Not only did the water becomedeeper, but oxygen levels also fell. Ironpyrites, commonly called ‘fool’s gold’,a mineral composed of iron and sulphurwhich forms only in the absence ofoxygen, is seen abundantly in rocks ofthe Mazzin Member. Lack of oxygen,‘anoxia’, means that nothing can live insuch waters (even the meanest of seacreatures requires oxygen to survive).

Similar successions are found in theSosio Valley in Sicily. However, neither

of these European sections seemedappropriate as an international markerfor the base of the Triassic since somecritical fossil groups were missing, andit was not possible to tell how completethe sections were. There appeared to besimilar gaps, perhaps even greater, inmuch of North America, where the LatePermian appeared to be largely absent.During the 1960s, stratigraphers turnedtheir sights on Asia.

Geotourism in Asia

In 1961, two American geologists, CurtTeichert and Bernhard Kummel, decidedto grasp the problem of the Permo-

Triassic boundary seriously. They readall the papers that had been publishedand decided they should visit everyavailable section. They identified keysites in southern China, Kashmir,northern Pakistan, the northern borderbetween Iran, Armenia and Azerbaijan,and northeast Greenland, and theyvisited all, or nearly all, of them, in animpressive example of geotourism witha purpose.11 Sadly, however, they couldnot visit southern China, largely forpolitical reasons.

In the Iranian-Armenian-Azerbaijanirock successions, Teichert and Kummelencountered many problems. Not least ofthese was that geologists had beenarguing for years over the exact location

of the boundary bed. They found thesame situation, if not even worse, in theGuryul Ravine in Kashmir, lyingbetween India and Pakistan. Since 1909,palaeontologists had confidently placedthe Permo-Triassic boundary at no fewert h a n seven different levels, rangingthrough a rock thickness of 400 metres inall. With such confusion and dispute,Kummel and Teichert had to sift throughmasses of contradictory evidence. Theyopted for the lowest position of all,below the first appearance of thebivalve Claraia. However, gaps in thesuccession of rocks suggested to themthat the Guryul Ravine section would notmake an ideal global stratotype.

On to Pakistan. Kummel and

Teichert visited the classic sections inthe Salt Range mountains of northernPakistan. These had been identified as arichly fossiliferous Permo-Triassicsuccession in the nineteenth century bythe German ammonite expert W. Waagenand others. In the 1950s, OttoSchindewolf and his students returned tothe Salt Range, in their efforts toestablish what had happened during theend-Permian mass extinction, at a timewhen his ‘neocatastrophist’ views weredistinctly unfashionable (see Chapter 4).

The Salt Range sections wererevisited in 1991 by two Britishsedimentologists, Paul Wignall fromLeeds University and Tony Hallam fromthe University of Birmingham. They

initially believed that there was a majorgap in sedimentation just below thePermo-Triassic boundary in the SaltRange, but they reversed their view afterfurther study of the conodonts.12 Thelatest Permian Chhidru Formationconsists of sandy limestones containingabundant fossils – bellerophontidgastropods, bryozoans, foraminiferans,crinoids, echinoids, brachiopods andalgae – evidence for a rich and diverseshallow marine fauna.

The overlying Kathwai Member,which spans the Permo-Triassicboundary, begins with sandy limestonesthat contain Permian brachiopods andbellerophontids, while higher up there isa switch to thinly bedded sandy

limestones, probably indicating deeperwater. The main extinction levelcoincides with this switch insedimentation style. The overlyingMittiwali Member is also generallythinly bedded, and black shales maketheir appearance in places. Wignall andHallam interpret this as evidence fordeepening of the water and a reductionin oxygen levels.

So, Wignall and Hallam hadconfirmed a switch from thick limestoneunits in the latest Permian to more thinlybedded limestones and black mudstonesin the earliest Triassic. But, faced withproblems of identifying exactly wherethe Permo-Triassic boundary lies in theSalt Range, it has been harder to

pinpoint the precise pattern of theextinction. If their interpretation of theconodont evidence is correct, it wouldmean that the end-Permian massextinction actually happened in theearliest Triassic in Pakistan.

The golden spike is driven

The best evidence for events at thePermo-Triassic boundary seems to comefrom the north shore of the greatPalaeotethys Ocean (Fig. 24). Thesections in northern Italy, in Iran-Armenia-Azerbaijan, in Pakistan, inKashmir and in south China all lie in thisvast tract of land. And the boundary

sections, from Italy to Kashmir, allappear to show a switch from normallimestone beds in the latest Permian tothin-bedded pyritic limestones at theboundary.

The studies by Kummel andTeichert, and by other geologists, inthese Asiatic sections had been devotedmainly to sorting out the stratigraphy,and not much can yet be said about theexact pattern of what happened. Surelygeologists must be queuing up to do suchimportant studies? Sadly not at themoment. As Doug Erwin pointed out inhis 1993 book,13 most of these areascoincidentally lie in the midst ofterritorial disputes and civil wars. Thedifficulty of access of some of the

classic central Asian sections means thatwe have to wait for essential studies –detailed bed-by-bed sampling to recordhow the different fossil groups died out,allied with fine-scale dating andgeochemical studies.

Kummel and Teichert had beenunable to visit China during their globaltour of the Permo-Triassic boundary in1961. Here again politics held upprogress, since fieldwork wasimpossible for anyone in China duringthe Cultural Revolution (1966–76) and ittook some time for Chinese geologists torecover their activities. Liberalization ofthe political regime in China since 1980has now allowed proper scientificexchanges, and intensive fieldwork has

been completed on several Chinese rocksuccessions that span the Permo-Triassicboundary.

24 The Late Permian world,showing the distribution of thecontinents and sites of importantrock successions that span thePermo-Triassic boundary.

Finally, after much lobbying and

campaigning, the section at Meishan, inZhejiang Province, south China, wasselected as the global stratotype for thebase of the Triassic in 2000, with thegolden spike driven in at the firstappearance of the conodont Hindeodusparvus.14 Not only is the top of thePermian defined in China, but so too arethe other subdivisions of the LatePermian (Fig. 23), because of the goodquality of the marine successions there.So, 159 years after Murchison named thePermian, the limits of the system havefinally been pinned down.

The Meishan section

The Meishan section is exposed in fivequarries, all close together. Combininginformation from the different quarriesgives a section some 50 metres thick,with the critical boundary bed occurring40 metres above the base (Fig. 25). Thelower part of the succession, termed theBaoqing Member, consists of thick unitsof limestone, with thinly beddedlimestones in between. These limestoneswere laid down on a marine slope, asshown by evidence of movement undergravity and abundant fossils such asforaminiferans, brachiopods andconodonts. Rarer fossils includecephalopods (coiled molluscs),echinoderms (sea urchins and starfish)and ostracods (small crustaceans), again

all typical of shallow seas.The sediments tell the story of the

environments in which they weredeposited, according to Paul Wignalland Tony Hallam.15 The thick and thinlimestones attest to warm-water shallowseas, with some water currents thatwashed the shells and other animalremains around on the sea floor beforethey were finally incorporated into therock. The thin beds include some blackmuddy limestones, the black colourcoming from organic matter that remainsin the sediment. This indicates anoxic(‘no-oxygen’) conditions, since organicmatter is usually completelydecomposed by scavenging organismswhich can only survive if oxygen is

present. The thin-bedded limestoneslying between the black mudstones donot contain any burrows, furtherevidence of an absence of life on the seafloor, since there are always worms andshellfish that creep about on the sand andburrow into it in search of food orsafety.

25 The Meishan section, includingthe critical Permo-Triassicboundary horizon, as defined byinternational agreement in 2000.Extinction levels are marked witharrows (1, 2, 3).

On top of the sediments of theBaoqing Member are those of theMeishan Member, the last to bedeposited in the Permian. The rocks arelimestones again, and depositionalconditions were apparently similar tothose of the Baoqing. Near the top, thereis extensive burrowing in the limestones,indicating a return to conditions of fulloxygenation. Suddenly, everythingchanges. The thick, burrowed limestoneshave gone, and so too have the abundant

fossils.The highest limestone of the Meishan

(bed 24) is followed by 29 centimetresof clays and more limestone. First in theclays comes a pale-coloured ash andclay bed, then a dark organic-richmudstone, followed by a muddylimestone. In the Chinese system theseare numbered as beds 25, 26 and 27respectively. Above bed 27 comes along succession of thin limestones andblack shales, termed collectively theChinglung Formation, containing onlyrare, small burrows. Here is a majorsuccession of low-oxygen beds spanningone or two million years. Layers ofpyrite crystals scattered abundantlythroughout the Chinglung sequence

provide further confirmation of the low-oxygen conditions. Beds 25, 26 and 27are where it was all happening, so let’slook at them in more detail.

The boundary beds

In the Meishan quarries, the rocks thatmark the biggest mass extinction of alltime do not look particularly unusual,just a succession of grey and blacklimestones and mudstones (Fig. 26). Buta millimetre-by-millimetre dissectiontells a strange story.

26 Photograph of the Permo-Triassic boundary in the Meishansection, northeastern China. Beds24 to 28 are numbered. The end-Permian mass extinction happenedin three closely spaced phaseshere, at the base and the top of bed24, and at the top of bed 27. Theglobal stratotype for the base ofthe Triassic is defined in the middleof bed 27, with the first appearanceof the conodont Hindeodus parvus.

The last typical unit of the Permian,the limestone bed number 24 of theMeishan Member, contains Permianfossils, highly worn specimens ofbrachiopods and foraminiferans. At thetop of bed 24 is a mineral-rich layer ofpyrite and gypsum, the pyrite certainly

suggesting anoxic conditions. Themeaning of the gypsum has been debated.Gypsum is normally found as a salt thathas been produced by evaporation of seawater, but Wignall and Hallam havesuggested that here it may simplyindicate an interaction of the pyrite andlimestone during modern-dayweathering.

The so-called lower boundary clay,bed 25, is a thin band of pale-colouredclay, only 5 centimetres thick, whichcontains scarce Permian foraminiferansand conodonts. Under the microscope,this clay contains small iron-rich pelletsand decayed pieces of quartz thatindicate that it was modified from anacidic tuff, an amalgam of volcanic

fragments and ash from an explosivevolcanic eruption.

The upper boundary layer, bed 26,consists of 7 centimetres of dark,organic-rich limey mudstone whichcontains a mixture of Permian andTriassic fossils – Permian brachiopodsand goniatites and Triassic bivalves(clams) and ammonoids. Goniatites andammonoids are cephalopod molluscs,distant relatives of the modern squid andoctopus. Conditions during deposition ofbed 26 were low in oxygen, but notanoxic, based on the relatively diversefossils, and on geochemical evidence.

These two mudstone beds, 25 and26, form a distinctive black-on-whitemarker band, called an ash band. It has

been detected so far in 12 provincesthroughout China and is useful forgeologists wishing to make correlationsfrom location to location. The doubledeposition event that produced it musthave occurred over as much as 1 millionsquare kilometres of China. What kind ofprocess could have produced such a thinash band carpeting such a huge area?

Bed 27, a 17-centimetre thicklimestone, contains pyrite crystals hereand there, but it is also full of burrows,showing that bottom conditions were notparticularly low in oxygen. The unitcontains rare Permian brachiopodfossils near the base, in the so-calledunits 27a and 27b. Then these disappear,and the conodont Hindeodus parvus is

found for the first time at the base ofsubdivision 27c. Here, 5 centimetresabove the base of bed 27, Chinesegeologists have now convinced theworld that the golden spike that marksthe base of the Triassic should beplaced.

Dating the end of thePermian

Older geological time charts give agesfor the Permo-Triassic boundaryanywhere between 225 and 250 millionyears ago, while publications since 1980have homed in on dates of 245, 248 or

250 million years ago. These dates werenot very precise, however, being merelyinterpolations. Radiometric dates wereavailable for the middle of the Triassic(238 myr.) and for the base of theArtinskian in the Lower Permian (268myr.), and something from 245 to 250myr. sounded about right for the Permo-Triassic boundary. This was a broadguess, and the span of 30 million yearsbetween the two radiometrically datedfixed points was far too long.

The Chinese sections offeredfantastic new opportunities for dating –not only could the boundaries bedetermined by means of the fossils, butthere were also clay/ash bands in bed25, close to the boundary, and further

clay bands both below and above thePermo-Triassic boundary and thesecould be dated scientifically. Geologistsdescended on the section, and bags ofclay were whisked off to laboratories inChina and elsewhere in the world.

The first dates, published in 1991and 1992,16 were assessed using theuranium-lead method, and these gavemeasures of 250 ± 6 myr. and 251.1 ±3.4 myr. The uranium-lead method isbased on the change of the uranium-238isotope to lead-206, a transition with ahalf-life of 4500 million years. Theplus-or-minus (±) figure is theassessment of experimental error – therange of ages that were found byrepeated analyses in the laboratory. It is

not a measure of global error, in thesense that no one knows what themaximum range of possible ageestimates might be. Repeated agedeterminations by different laboratories,and using different isotope series anddifferent equipment, are the best test ofthe accuracy of any publishedradiometric date.

Such a test came in 1995, when anAmerican group17 used the new argon-39 to argon-40 technique to achieve adate of 249.9 ± 1.5 myr., well within therange indicated by the uranium-leaddates. In a further, even more detailedstudy, an American-Chinese team led bySam Bowring of MIT18 reverted to theuranium/lead method, but they used two

isotopic series, the decay of uranium-238 to lead-206, and the decay ofuranium-235 to lead-207, which has ashorter half-life of 700 million years. Byusing the two isotope series, this teamwas effectively cross-checking everymeasurement they made. The ash bandsin bed 25 were dated at 251.4 ± 0.3 myr.and bed 28, immediately above thePermo-Triassic boundary, yielded a dateof 250.7 ± 0.3 myr. So, beds 26 and 27,the latter of which contains the officialbase of the Triassic, fall between thosetwo dates, representing up to 700,000years (0.7 myr.). Splitting the differencegave a date for the Permo-Triassicboundary of 251.0 myr.

Despite the thoroughness of their

study, it has been claimed that Bowringand colleagues had not allowed forerrors coming from unaccounted lossand inheritance of lead, and from mixingof grains of slightly different ages withintheir samples. In 2001, Roland Mundilfrom the Berkeley GeochronologyCenter and colleagues refined theanalyses, and claimed they had to add 2myr. to the Bowring dates, giving avalue of 253 myr. for the Permo-Triassicboundary. However, we will use thedate of 251 myr. until the Mundilrevision has been tested and debated.The three apparent pulses of extinction,at the bottom and top of bed 24, and atthe top of bed 27, may, in total, spanabout 1 million years. But what was

going on in this relatively short interval?

Carbon isotope shifts

Isotope geochemistry can provideevidence about ancient environments aswell as rock dates. In the case ofenvironments, geochemists focus onisotopes of carbon and oxygen. Carbonisotopes show a sharp negativeexcursion (see Fig. 27), dropping from avalue of +2 or +4 parts per thousand to -2 parts per thousand in the pale-colouredmudstone, bed 25, which corresponds tothe main extinction level. What does thismean?

The impressive-looking term δ13C isessentially the ratio of the two stableisotopes of carbon, 13C and 12C (the 12and 13 are the atomic weights) measuredagainst a standard, and calibrated asparts per thousand. When plantsphotosynthesize, they take up the isotope12C from the soil (land plants) or fromseawater (the floating phytoplankton) bypreference, and this has the effect ofincreasing the proportion of 13C leftbehind. A geologist in the future whomeasures the δ13C ratio from the soils orsea-bed sediments being deposited todaywill note relatively high values, and willinterpret these as an indication of highlevels of biological activity, sometimestermed ‘productivity’.

A negative shift of 4 to 6 parts perthousand in the δ13C ratio might seemfairly minor, maybe just a local effect.This negative excursion, however,seems to be a global phenomenon,having been found in Permo-Triassicsections worldwide. It is also actuallylarge. At the KT boundary, the negativeexcursion is smaller, a mere 2 or 3 partsper thousand. The drop in the δ13C ratioat the Permo-Triassic boundary could beinterpreted as evidence for a majorreduction in biological productivity, justas would be expected at a time of massextinction. But, as we will see later,even the most astonishing level ofextinction could not produce such adrop.

27 The extinction of life at the endof the Permian in China, showingradiometric ages, the carbonisotope curve, and the ranges of333 species of fossils identifiedfrom 90 metres of rock in theMeishan quarries. A, B and C markthe three apparent pulses ofextinction.

Oxygen isotopes in limestones show

a similar negative shift, from a highvalue of −3 to −1 parts per thousand inthe Late Permian, to −7 just at theboundary. The oxygen isotope ratio,δ18O, is often used as apalaeothermometer, and a drop of 4parts per thousand could indicate aglobal increase in temperature of around16°C.

So, in summary, the carbon isotopescould suggest a major drop inproductivity (but there must be muchmore to it than simply that), while theoxygen isotopes suggest thattemperatures rose dramatically at thesame time. These are the simplestinterpretations of the geochemical data,however, and there are many

complicating factors that we mustexplore later in trying to work outexactly what happened to the world 251million years ago (see Chapter 11). Wehave seen evidence of some startlingchanges in marine environments in theMeishan section. But what of life inthem?

The biggest extinctionanatomized

Chinese and western palaeontologistshave collected fossils intensivelythrough the Meishan section in anattempt to discover exactly what

happened. Early reports suggested thatthere had been three levels of extinction,perhaps spanning half a million yearsaltogether. This clearly had to beassessed in the greatest detail possible,so a Chinese-American team set aboutthe task. Y. G. Jin and his colleaguesfrom the Nanjing Institute of Geologyand Palaeontology, and Doug Erwinfrom the National Museum of NaturalHistory in Washington, undertook a hugesampling programme.19

They collected fossils from 64levels through the 90 metres of rocks thatspan the Permo-Triassic boundary in theMeishan quarries, from bed 1 to bed 46of the standard numbered succession. Inall, 333 species were identified,

belonging to 15 marine fossil groups –microscopic foraminifera, fusulinids andradiolaria (all microscopic organisms,some of them living on the sea floor andothers floating at the sea surface), rugosecorals and bryozoans (both of whichform colonies made from numerousindividual animals living in a sharedstructure), as well as brachiopods,bivalves, cephalopods and gastropods(shellfish that lived on, or near, the seafloor), ostracods (swimming, shelledshrimp-like creatures), trilobites,conodonts, fishes and algae (seaweeds)– and they plotted their exact rangesagainst the stratigraphic section (Fig.27).

The plot shows a number of minor

extinction levels low in the section. Inall, 161 species became extinct belowthe boundary beds (levels 24 to 27),during the 4 million years before the endof the Permian. Extinction rates inparticular beds amounted to 33% orless. Then, just below the Permo-Triassic boundary, at the contact of beds24 and 25, most of the remaining speciesdisappeared, giving a colossal rate ofloss of 94% at that level. But was therea single pulse of mass extinction, orthree?

The Signor-Lipps effect andforward smearing

It is important to determine whether theend-Permian mass extinction happenedat one level – in a geological instant – orwhether it occurred in several pulses.Does the Meishan section show a singleextinction event, the main step in thefossil ranges at 251 myr. (Fig. 27: levelB), or three levels of extinction (A, B,C)? Even though the extinction steps Aand C are much smaller than B, they arestill there. Palaeontologists, however,are always cautious about reading therocks literally.

The rocks may mislead the unwaryinvestigator. Imagine that extinction ratesare constant, with a speciesdisappearing, on average, every 100,000years. If sedimentation is not relatively

constant, the regular backgroundextinction pattern may be disrupted. Agap in sedimentation, representing half-a-million years would automaticallycreate an apparent sudden extinction offive species. But that’s not all.

Palaeontologists may be sure of onething: they will never find the lastsurviving member of a particularspecies. This observation is based onthe low probability that any individualorganism will actually become a fossil,and the likelihood that the last stragglersof a species may be holed up in someobscure location that never fossilizes, oris never found. Therefore, when a massextinction occurs, and 100 species dieout in an instant, the best a

palaeontologist can hope for, even aftermonths of assiduous rock smashing, is apattern where only 20 or so actuallydisappear in the extinction level, and theremaining 80 seem to drop out in thelevels below. So, palaeontologistsexpect to find backwards smearing of amass extinction event, a superficiallygradual dropping out of species one-by-one. This is the Signor-Lipps effect,named after the Americanpalaeontologists, Philip Signor and JereLipps, who first presented the idea instatistical terms.20

Smearing can occur forwards aswell. Fossils may be found by chanceabove their extinction level. Thishappens commonly when sediments are

burrowed by worms and shrimps, whichcan churn the sands and muds to a depthof a metre. In a section like Meishan, ametre of rock in the Permo-Triassicboundary beds can represent severalhundred thousand years. Fossils in theburrowed rock may be moved up anddown by the burrowers. Careful study ofthe rocks should show whether fossilsare in place, or whether they have beenmoved by burrowers, but upwardssmearing of a mass dying level has to beconsidered.

In their 2000 paper, Jin andcolleagues assessed the reliability ofextinction levels A and C (Fig. 27).Could they simply be backwards andforwards smearing of a single mass

extinction, level B, at 251.4 millionyears ago? Steps A and C are certainlymuch smaller than B, which immediatelysuggests the possibility that they do notmark real events. Following their carefulcalculations, the investigators found thatthe lower level A simply disappeared:the six species that apparently died outat this level are almost certainly falselast records, evidence of the Signor-Lipps effect.

As for Level C , there is evidencefor burrowing in bed 27, for example,but the fossils have not been shifted upthrough 17 centimetres of limestonehere. About 45 species apparentlydisappear between levels B and C (17 atlevel C), and further species drop out in

levels above. The sedimentologicalobservations, and the statistical tests, donot suggest that these have all beensmeared forwards from the true massextinction level. The picture is of asingle mass dying at level B, and then aperiod of about 1 million years duringwhich surviving species died out. Aseparate event C cannot really be pickedout – it is part of a run of extinctionsafter the main event.

Is this, however, merely a localpattern, found at Meishan, but notelsewhere? If it is, then it cannot tell usso much about what happened 251million years ago. But, if the samepattern of geochemical changes, anddying, can be found all over the world,

then geologists can begin to look for theglobal-scale processes behind the end-Permian mass extinction.

Events in China andPakistan

In 1991, after visiting Meishan, PaulWignall and Tony Hallam went on toexamine other sites in south China.21 TheHushan section, 200 kilometres to thenorth of Meishan in Anhui Province,‘records a virtually identical series ofearliest Triassic events to those seen atMeishan’. There are the white and blackclay bands marking the boundary, and

containing a mixture of Permian andTriassic fossils. The same chemicalchanges are documented, with a majordrop in oceanic productivity, a rise intemperature and a progressive shift toanoxic conditions in the earliestTriassic. Numerous other Chinesesections show the same pattern, and theextinction event seems to happen at thesame time in each section.

In Pakistan, however, as we haveseen, the end-Permian mass extinctionmay have happened rather later, in theearliest Triassic. This possibledifference in timing of the massextinction in China and Pakistan is farfrom certain. Admittedly, south Chinaand northern Pakistan are some distance

apart today – about 4000 kilometres –and they were probably as far apart inthe Late Permian. At that time (see Fig.24), both regions were under the sea, asthe sediments show, Pakistan on thenorth shore of the great Tethys Ocean,and China part of a microcontinent to theeast.

The sedimentary successions inPakistan and China accumulated on theshores of separate continents, but adifference in timing of the massextinction would lead to a very differentmodel for what happened 251 millionyears ago from one in which events wereexactly coincident. What of the rest ofthe world?

The Panthalassa Ocean

Not much is known about what happenedin the vast Panthalassa Ocean. Thisocean (see Fig. 24), surrounding thecontinents, corresponds to the modernPacific, but it was much wider, since theAtlantic did not exist. In the Mino-TanbaBelt in southwest Japan, on the westshore of the Panthalassa Ocean, theboundary rocks consist of thicknesses ofred cherts: hard silica-rich, almostglassy rocks, made up from the skeletonsof radiolaria, minute planktonic animalsthat fell to the sea floor and accumulatedslowly for millions of years. Themonotony of these cherts is broken bythe Toishi Shale that spans the Permo-

Triassic boundary. Exactly at theboundary is a black shale. Why thechange in sedimentation?

A red colour in sediments indicatesthe presence of abundant oxygen, just asa black colour usually indicates anoxia,the absence of oxygen. Grey issomething in between. Colours are auseful diagnostic tool for geologists, andthey are usually reliable since thecolours are produced by the sameproperties of the rocks in most cases.Red colour in a sandstone, a mudstoneor a chert, often indicates the presenceof iron oxide in the form of haematite. Ifhaematite is deposited in a sedimentaryrock in low-oxygen conditions, much ofthe oxygen is lost, and the haematite

transforms into other iron minerals thatare grey or green in colour. A blackcolour usually means that there is a greatdeal of organic carbon – the remains ofdead plants and animals which have notbeen scavenged by detritus-eaters orcombined with oxygen to form carbondioxide.

Simply interpreted, then, theJapanese section shows well-oxygenatedmarine conditions in the Late Permian(red cherts), after which oxygen levelsfell (grey cherts), falling further to avery low level in the lower Toishi Shale(grey shales), to almost none at all at theboundary (black shales). Then, in aprecise reversal of the sequence, oxygenlevels built up in the Early and Middle

Triassic, a span of 15 million years,from black shales, to grey shales, to greycherts and finally to red cherts again.

The Japanese section is not yet wellenough dated for palaeontologists to besure exactly how the end-Permianextinctions happened. Late Permianradiolarians died out as oxygen levelsfell, but that is all we know. Whether theJapanese extinctions happened at thesame time as the main mass extinctionhorizon in China, or before or after it,cannot yet be ascertained. Events innorthern continents may have been rathermore drawn out, however.

Greenland

Geologists, from Kummel and Teichertonwards, have also turned their eyesnorthwards, to Greenland. In the LatePermian, a narrow seaway extendeddown from the great northern BorealOcean into Laurasia (see Fig. 24),bringing shallow marine sedimentationand faunas into the northeastern coast ofthat continent.

The Greenland succession consistsof a series of rocks called the SchuchertDal Formation, essentially latestPermian in age, overlain by the WordieCreek Formation, dated as Triassic. TheSchuchert Dal Formation consists of amixture of marine limestones, shales and

gypsum containing many fossils and it isheavily burrowed. The Wordie CreekFormation is composed of dark greysandstones and siltstones, some of themcontaining pyrite. Fossils are rare,especially in the lower portions, andthere is little sign of burrowing. So herewe see the usual switch from limestonesand mixed sediments with rich fossilfaunas to deeper-water anoxicmudstones, with little sign of life, afterthe end-Permian event. However, thefossils in the Greenland successionscontain a conundrum.

Examples of Permian fossils –brachiopods, corals, foraminifera,crinoids, echinoids and bryozoans – arefound as much as 20 metres into the

‘Triassic’ Wordie Creek Formation.One explanation could be that the finalextinction of Permian organismshappened slightly later in the BorealOcean than in equatorial regions. CurtTeichert and Bernhard Kummel couldnot accept that. Surely the Permianfossils should have disappeared at theend of the Permian, in other words, atthe top of the Schuchert Dal Formation?But how then could one account for thepresence of abundant, and undamaged,Permian fossils so high in the WordieCreek Formation?

Armoured mudballs.22 Teichert andKummel argued that the delicate Permianfossils in the Wordie Creek Formationhad been reworked in the Early Triassic.

The beasts had indeed died at the end ofthe Permian, leaving their skeletons inthe Schuchert Dal Formation. Erosion byseabed currents in Wordie Creek timessupposedly cut into the Schuchert Dalbeds, releasing the shells and otherremains. But how could these remainshave been transported without beingdamaged? Teichert and Kummelsuggested that great balls of mud hadpassed over the older rocks, picking updelicate shells and fragments on theirsurfaces, and that these had then come torest, dissolved somehow, and had leftthe older fossils in younger sediments.

Hallam and Wignall gave theirsuccinct view of ‘[t]his franklyridiculous scenario’. It was clearly

nonsense, they suggested, since noexamples of giant dissolving mudballshave ever been reported, and thePermian-style fossils in the first 20metres of the Wordie Creek Formationare preserved more or less in lifeposition. In other words, the shells aresitting on their backs as they would havewhen alive; the corals and bryozoansstand upright like little trees. No rollingmudball could do that.

So were extinctions in Greenlandlater than in China? Was there alatitudinal, perhaps climatic, effect?Probably not.

Timing of the end-Permianevent

The clue to timing is in the Meishansection. There the end-Permian massextinction happened essentially in ageological instant, but the initial eventwas followed by a period of thousandsof years when species continued todisappear (see Fig. 27). Perhaps theseemingly later extinctions in Pakistanand Greenland correspond to part of thislater decline. It is striking that the highlycondensed sections in China (some 50metres document 6 myr.) show the samepatterns as the extended Greenlandsections, where the equivalent time-span

is represented by over 500 metres ofsediments.

New work, published in 2001, hasgiven greater insights into the timing ofevents. Richard Twitchett, then based atthe University of Southern California,and Cindy Looy from the University ofUtrecht, and their colleagues,23 havebeen studying the Greenland sections inintense detail. They find that in fact themarine record there documents acollapse of ecosystems within 50centimetres of the section. At the top ofthe Schuchert Dal Formation, life wasnormal: a typical latest Permianassemblage of brachiopods, corals,ammonoids and foraminifera ispreserved. The sediments are intensely

burrowed, indicating a well-oxygenatedenvironment.

Then, through the next 50 centimetresof green and grey muddy siltstones, lifeis devastated. The burrows disappearand there are almost no fossils. All theseabed life that existed just a fewcentimetres below has gone, and only afew species reappear above theboundary beds (the late survivors atMeishan and the ‘armoured mudball’fauna in Greenland), but these Permiansurvivors rapidly disappear. At the baseof the Wordie Creek Formation areblack anoxic mudstones containingvirtually no fossils of any kind. Based ontheir calculations of rates of deposition,Twitchett and Looy estimate that the 50-

centimetre death bed representssomething between 10,000 and 60,000years.

It is too soon, however, to be surewhether this time-scale is trueworldwide. Only the Meishan and theGreenland sections have received thesort of detailed stratigraphic study that isnecessary to plot the precise course ofevents. Until geologists andpalaeontologists return to the sectionsoutside China and Greenland to collectthe fossils centimetre-by-centimetre, andobtain some good radiometric datesusing modern analytical machines, weshall be uncertain.

And until such detailed results areavailable, it will be most sensible to

read the patterns from China andGreenland as indications of whathappened. We have seen that diversegroups of brachiopods, corals,bryozoans, foraminiferans and othersdied out. It is important now to look atthose groups of marine animals to try tounderstand the seriousness of the event.Ecosystems were ripped to shreds, butwhat were those ecosystems?

8

LIFE’S BIGGESTCHALLENGE

Doug Erwin of the SmithsonianInstitution in Washington called the end-Permian event ‘the mother of all massextinctions’, a reference to the remarkmade by Saddam Hussein, President ofIraq, in 1990 that the Gulf War would be‘the mother of all battles’. Erwin

characterized the biotic effects of theextinction:1

Killing over 90% of the species in the oceansand about 70% of vertebrate families on land isremarkably difficult. The end-Permian massextinction was the closest metazoans [animals]have come to being exterminated during thepast 600 million years. The effects of theextinction are with us still, for it changed thestructure and composition of marinecommunities far more than any event since theCambrian radiation.

This is dramatic stuff, but noexaggeration. The quotation highlightsthree topics that we will consider in thischapter.

First, it is necessary to review thedifferent groups of animals in the sea,

how they lived and their role inecosystems in the Late Permian, justbefore the event. In each case, it isimportant then to determine whathappened to each group. Did they all dieout at once, or were some more resistantto extinction than others?

Second, it will be critical to estimatejust what was the magnitude of the event.Palaeontologists bandy all kinds offigures around for the percentage ofspecies that were lost: 75%, 80%, 90%,or 96%. In a sense, it doesn’t reallymatter, since such losses mean that only4–25% of species survived, a tinyfraction of the total. Equally, however, itis important to get the figures right, forcomparison with the KT event and

others of that magnitude – and indeedwith present-day rates of species loss.

Third, it is necessary to try tovisualize just which plants and animalsmanaged to creep through the crisis.Which groups survived, and which wereentirely wiped out? The broader-scalerecovery after the end-Permian eventwill be considered later (Chapter 11),but death and survival will beconsidered here. In addition todetermining the victims and thesurvivors, it is important then to enquirewhether there is any evidence forselectivity. Did the survivors share anyfeatures that set them apart from thevictims?

Before and after

Comparison of the scene before andafter the end-Permian event illustratesthe severity of the mass extinction (Figs28, 29). The ‘before’ scenes, from thelatest Permian, show reef communitieswith rich faunas of corals, bryozoansand crinoids, with ammonoids and fishesswimming above. Brachiopods, seaurchins, snails and foraminifera rest, ormove slowly, on the bottom. These arerich and complex ecosystems, whetheron the shores of the Tethys Ocean, theChinese example, or on the edge of theBoreal Ocean, the Greenland example.

In the earliest Triassic the reefs havegone, and all the organisms that were

specialized to form reef structures havedied out. The rich accumulation of shellydebris is missing, since everything hasdied, and each scene is dominated by asingle kind of bivalve, Claraia. Theearliest Triassic Greenland fauna seemsto be richer than that of China, with asmall shark and other small fishes, acouple of ammonoids and a conodontanimal swimming over a sparse fauna ofClaraia and some small snails.

These impressions of life before andafter the great extinction show just howsevere it was. A global survey of themajor groups of organisms in the seaconfirms the enormity of the changes.

Microscopic floaters

Some of the key organisms in the sea areinvisible to the naked eye. These are theplankton, the microscopic plant-like andanimal-like organisms that float near thesurface. The plant-like plankton captureenergy from the sun by photosynthesis,just as green plants do on land, and theyare fed on by the animal-like plankton.Planktonic organisms keep afloat by anamazing array of special devices –broad flanges, spikes, gas-filledballoons – that stop them from sinking;some have spiral body shapes so thatthey spin slowly. They include manyunique groups that spend their entirelives in that form, while others are the

larvae of typical marine animals, such ascrabs, sea urchins or corals, that willeventually metamorphose into their adultforms. The plankton form the base of allfood chains in the sea. They are eaten byshrimps, fishes and other larger animals,and these in turn form the diet of largerfishes, sharks, seals and whales. Kill theplankton, and you kill all life in the sea.

28 Before and after: the tropicalocean. A reconstruction of thelatest Permian sea-bed (above) andthe earliest Triassic sea-bed(below), immediately after thecatastrophe; based on informationfrom the Meishan section, China.

29 Before and after: the northernArctic ocean. A reconstruction of atypical latest Permian sea-bed(above) and an earliest Triassicsea-bed (below), immediately afterthe catastrophe; based oninformation from Jameson Land,East Greenland.

The radiolarians, delicate net-likelittle organisms with a light skeletongenerally made of silica (silicondioxide, the main component of sand andof flint), today feed on bacteria andplant-like plankton. Their skeleton ismade up of tiny spicules, or needles, ofsilica, forming perforated spheres, somewith spikes, just like miniatureChristmas decorations. Others are like

tiny string shopping bags, suspendedfrom a single corner. All are perforatedwith numerous regular holes. When theydie, the flinty little skeletons ofradiolarians rain down on to the deepocean floor where they accumulateslowly, at only 4 or 5 millimetres per1000 years. Nevertheless, over millionsof years, radiolarian oozes havesolidified into cherts, glassy pure-silicadeposits, that currently make up about3% of the modern ocean floor.

In Late Permian deep marine rocksuccessions, radiolarian cherts are foundquite commonly in China, Japan andCanada, but then disappear completelyat the end of the Permian, only toreappear in the Mid Triassic, some 8

million years later. This ‘chert gap’ ismatched by the virtual annihilation of allspecies of radiolarians at the end of thePermian – an event that is particularlystriking since otherwise the radiolarianshad had a singularly uneventful historyfor the previous 450 million years. Ittakes some enormous environmentalshock to kill off such minute planktonicorganisms that must have been present asmillions of millions of individuals allover the world. Far easier to kill offlarge and less numerous animals such asdinosaurs.

Another element of the planktontoday are the foraminifera. Foraminiferalook like tiny spiral or coiled snails,with a shell made from calcium

carbonate (calcite) or glued sand grains,enclosing their single-celled soft parts.The shell may form a tall spiral, a flatcoil or disc, or a mass of small globules,in each case being divided into a numberof internal chambers. In the Permian allforaminifera were sea-bed dwellers,feeding themselves on the rain of organicmatter and plant-like plankton that sankfrom the surface. The dominantforaminiferan group in the Permian werethe fusulines. Their shells were madefrom many tiny crystals of calcite. Someof them reached as much as 10centimetres in length – most unusual forsuch a single-celled animal. Thefusulines flourished in the Permian,evolving fast and giving rise to over

5000 species. Indeed, they evolved sofast, and achieved such diversity, thatthey are used as important guide fossilsfor dating Permian marine rocks. Thenthey all disappeared.

Until recently, the die-off of thefusulines was thought to have lasted formost of the second half of the Permian,some 20 million years or more. It wassaid to have been a gradual die-off,perhaps linked to long-term climaticdeterioration. But new work has shownthat the early studies had been toolimited, and had relied on a literalreading of the record. The problem wasbackward smearing, the Signor-Lippseffect, as noted above (Chapter 7).

A recent study of the Permo-Triassic

boundary in Austria2 by MichaelRampino and Andre Adler, both of NewYork University, has shown on the basisof very detailed collecting that mostfusuline species did indeed go extinctright at the Permo-Triassic boundary.Other fusuline species disappeared fromthe rock record as much as 16 metresbelow the boundary, which could betaken to imply a long-term die-off.However, these species were rareforms, known only from small numbersof specimens. Rampino and Adlerargued that these were misleading datumpoints: the disappearances do notindicate extinctions. If these rare formshad been commoner, they wouldprobably have been sampled up to the

Permo-Triassic boundary as well. Asmentioned before, with a patchy fossilrecord, it is unlikely thatpalaeontologists will ever find the verylast member of a particular species tohave lived on Earth. If a hundred speciesdied out in an instant, the fossil recordmight still paint a picture of long-termgradual disappearances. Rare forms areespecially liable to this phenomenon offalse early disappearance, the Signor-Lipps effect.

Not all foraminifera died out duringthe end-Permian crisis. The generallylarge fusulines had completely crashedout of sight. But survivors includedmainly smaller forms and some flattenedspecies that burrowed in the sea-floor

sediments feeding on detritus. Acharacteristic feature of these, and othersurvivors, is that they may have beenadapted to living in conditions of lowoxygen. Perhaps this is a clue to thenature of the environmental stresses atthe end of the Permian.

Reefs

Reefs were common in Permian shallowtropical waters. For example, huge reefsdeveloped over much of west Texas andNew Mexico. During the Mid Permian,as the tropical seas became deeper,reefs built up around the edge of theancient Delaware Basin, reaching a

height of some 600 metres. Just likemodern coral reefs, the living corals andother animals, were in the upper parts ofthe reef, keeping in close touch with thesea surface which allowed theassociated plant-like organisms tophotosynthesize. Deeper parts of the reefwere formed from overgrown, deadcoral skeletons, as well as shells andother reef rubble.

Reef life then was hugely diverse,with hundreds of species living in closeproximity. The framework of the reefwas built from sponges, corals andbryozoans, animals that secrete a stonyskeleton in which they live. Living onthe dead corals were various clingingmolluscs and worms. Creeping among

the coral fronds were snail-likemolluscs, sea urchins, starfish andshrimps. And swimming above were jet-propelled nautiloids and ammonoids(relatives of squid and octopus),swimming arthropods and fishes ofvarious kinds. Just as today, LatePermian reefs were diversity hotspots –locations of unusual species richness.

The seas have retreated now ofcourse, but the west Texas landscape hasnot changed much in the past 250 millionyears. The visitor today can essentiallystand on the Late Permian sea bed andlook around at the towering GuadalupeMountains, made up of reef limestones,termed the Capitan LimestoneFormation. The vast size of the reef,

hundreds of metres thick and 400kilometres long, is immediately clear,and the richness of Mid Permian tropicalreef life is evident. Such large and richlydiverse reefs are also known from Chinaright to the end of the Permian.

Reefs were entirely wiped out by theend-Permian event. Like the ‘chert gap’,there is a ‘reef gap’ lasting for some 7 or8 million years in the Early Triassic.Many sponge groups survivedapparently unaffected through the end-Permian crisis, but others, especiallythose associated with the tropical-beltreefs, were decimated.

The corals were even harder hit.Throughout the preceding 200 millionyears, limestone deposits of tropical

zones are absolutely teeming with theskeletons of rugose and tabulate corals.Any novice fossil collector will haveaccumulated dozens of specimens ofthese corals – the tiny ice cream conesof small, solitary rugose corals and thefist-sized, rounded tabulate coralcolonies composed of dozens of regular,honeycomb-like or star-shapedchambers. Some are even shaped like abursting sun – hence the name of thecoral, Heliolites, the ‘sun rock’. Solitarycorals built themselves tubular housesfrom calcite which they laid down roundand round their soft bodies as protectionfrom predation. Their tubular housesrange in size from a few millimetreslong to the vast metre-long cones of

Caninia in the Carboniferous. Mostlythey were fixed upright on to rocks orother hard materials on the sea bed.

Colonies of rugose or tabulate coralsare among the most beautiful of fossils.Colonies arose when one progenitorcoral animal cemented its skeleton downto the seafloor, and then set aboutbuilding its stony dwelling chamber. Byendlessly splitting, the original coralanimal formed numerous identicalclones of itself, each of whichconstructed a little chamber. There iseconomy in such an arrangement, sinceeach subsequent coral animal has tobuild only a few side walls and can usethe pre-existing parts of the colony. Inthe end, most colonies formed bulbous,

cabbage-shaped structures or broadplates on gradually expanding trumpet-like stalks. Each coral animal kept itselffree of the others, and fed by capturingfood particles on sticky tentacles. Ifdanger threatened, the coral animalswithdrew deep into their stony houses.

The rugose and tabulate corals,which had been the mainstay of reefformation worldwide for 200 millionyears of the Palaeozoic, all died out atthe end of the Permian. It seems that theyhad undergone a long-term declinebefore the very end. First to go were themassive colonial forms, and at the end itwas the turn of the colonies made fromless intimately intergrown tubes and thesolitary forms. The early losses of some

coral groups seem to relate to changinghabitats. For example, the warm tropicalseas that had covered Texas and NewMexico had withdrawn in the LatePermian. This was not part of the crisis,merely a change in sea levels andcontinental positions. Where coral reefsare found in the latest Permian, however,such as in South China, the coralssurvived right to the end.

Other reef-builders are less wellknown. The bryozoans, or ‘mossanimals’, form regular colonies ofhouses which are generally tiny – lessthan a millimetre across – each for alittle individual animal. Often, bryozoancolonies are sheet-like structures thatgrow vertically from the seafloor, while

others encrust corals or shells as a finemeshwork. Of four major bryozoangroups around in the Permian, onebecame extinct during the crisis and theother three suffered heavy losses ofspecies.

After the ‘reef gap’, the first reefs toform in the Mid Triassic were smallpatches on the sea floor, known fromshallow-water sediments of theDolomites in north Italy. But it tookmany millions of years more beforelarge-scale framework reefs like thegreat Capitan Reef of the GuadalupeMountains in Texas reappeared. Thefirst Triassic reefs were formed fromsome surviving species of bryozoans,stony algae (distant relatives of

seaweeds), sponges and the firstmembers of a new group, thescleractinian corals. Scleractinian coralsdominate modern reefs, with theirmulticoloured fleshy bodies andtentacles enclosed within stonychambers, and they owe their origin tothe after-effects of the end-Permian massextinction.

Sea lily forests andbottlenecks

The echinoderms (‘spiny skins’) includesea urchins, starfish, sea cucumbers andsea lilies. They are all characterized

both by a skeleton made from calciteplates and their five-rayed symmetry,whether the five legs of a starfish (ten insome) or the ten-panelled structure of asea urchin. Echinoderms were hitparticularly hard by the end-Permiancrisis – indeed, most had disappearedseveral million years before.

Most of the fixed echinodermsdisappeared. Sea lilies (called moreproperly crinoids) were hugelysuccessful in the Palaeozoic, formingparts of the great reefs, either growingon and around the corals and sponges orforming huge crinoid forests on theirown. Typical crinoids look like plants: along flexible stalk fixed down by a root-like structure, with a blob at the top of

the stalk and tentacle-like armsshimmering in the currents above.Crinoids feed on small organic particlesin the water, which they capture on theirsticky arms and then waft down a centralgroove on the upper surface of the arminto the mouth, located in the centre ofthe body. The body is the blob on top ofthe stalk. Two subclasses of crinoidsdied out at the end of the Permian, andthe recovery of crinoids in the Triassicwas probably founded on a single genusthat survived through the crisis. Thereare still some stalked crinoids today, butmost are free-swimmers, rolling andcrawling along the sea floor, propelledby their slender, multicoloured,billowing arms.

Other stalked echinoderms that hadonce dominated the Palaeozoic seafloors were also wiped out. Prominentamong these were the blastoids, whichlooked like stone tulips or gooseberrieson sticks. Lower than most crinoids, theyformed large spiky forests around reefsin the Carboniferous and Permian. Theydisappeared without trace at the end ofthe Permian.

Both these losses effectivelydestroyed the very strange echinodermforests of the Palaeozoic sea floor.Catastrophic for the crinoids andblastoids of course, but also for all thelittle creatures that lived in, on andbelow the forests. In some ways, cuttingdown the echinoderm forests on the sea

floors was like cutting down a trueforest on land. Dozens of otherorganisms depend on the forest for theirlivelihood, and the same was true for thecrinoid-blastoid forests of theCarboniferous and Permian. Their lossmarked the end of a unique habitat thathas never been reconstructed on theocean floor.

Starfish and sea urchins were alsonearly wiped out. Indeed, the end-Permian crisis marked something of abottleneck for both groups. A bottleneckin evolutionary terms is a time ofextreme reduction in diversity of agroup. The group flourished before andafter the bottleneck, but what comes afterclearly can represent only a small

sample of the original population. So,post-Palaeozoic sea urchins and starfishwere quite different from theirPalaeozoic forebears. Among starfish,their most delightful specialization, theability to turn their stomachs inside-out,engulf their prey and begin digesting itbefore even eating it, arose only in theTriassic and Jurassic. Permian starfishhad to content themselves with a morediscrete and concealed stomach.

Only one genus of sea urchin,Miocidaris, is known to have survivedthe end-Permian crisis. This was anextreme bottleneck, cutting globaldiversity from 20 or 30 species down toone or two. From this one genus,seemingly, arose all the diversity of later

sea urchins. In the cases of starfish andsea urchins, the bottleneck actually had apositive effect. Both groups successfullymultiplied in numbers through theTriassic and rose to become importantcomponents of the seabed fauna, as theyare today. Of course, had this nothappened – if, for example, the tinynumbers of species that squeezedthrough the bottleneck from the Permianto the Triassic had actually all died out –the effects would have been negative.But then it would have been a totalwipeout, pure and simple.

Shellfish

First-year geology students alwayscomplain about having to learn thegroups of fossils. One of the key factsthey have to grasp is the differencebetween brachiopods and bivalves.These are two distinct groups of shelledanimals which have different ancestors,but which look superficially similar. Abrachiopod consists of two shell halves,more properly called valves, thatenclose and protect the animal inside.The shell is fixed to the seabed by atough thread that emerges from the tip ofone of the valves. The two valves arejoined along the hinge line, and they maybe opened by muscular activity to allowfood particles to be sucked in and wastematerial expelled.

It is easy to tell a brachiopod from abivalve: brachiopod valves are differentin dimensions, while those of bivalvesare identical mirror-images. One valveof the brachiopod is larger than theother, and it is often shaped like aRoman oil lamp – teardrop-shaped, withthe extension at the hinge-end oftenperforated by a large circular hole forpassage of the attachment thread (justlike the hole for the wick in the Romanoil lamp). The other valve is circularand smaller. Bivalve valves, on theother hand, generally fit exactly overeach other, being identical in size.

The brachiopods and the molluscs ofthe Permian were hit hard by the massextinction. Molluscs, such as clams,

oysters, mussels, whelks, octopus andsquid, dominate the seafloor today,while brachiopods are relatively rare,being found only in rather deep watersand confined to certain parts of theworld. However, the situation was thereverse in the Palaeozoic, and the end-Permian crisis perhaps has a large partto play in engineering the switchover.

To human eyes the brachiopods mayseem pretty limited in their potential –all they really do is sit on the seabedopening and closing their valves. Theyfeed by sucking water, plus foodparticles, into their shells, passing itover a looped filtering organ, thelophophore, and blowing water out theother side. However, the Permian was a

time of astonishing innovation in thegroup. The cone-shaped richthofenidscopied the corals, cementing themselvesto a rock or another shell with the tip ofthe cone and standing upright in tightclusters to form mini-reefs. The smallervalve had become simply a small lid,like that of a pedal bin, which could beopened to allow feeding. The fat, andoften large, brachiopods also flourishedin the Permian, when remarkable newspiny forms appeared. The spines weredelicate tubular structures sproutingwildly all over the base valve, and theseextraordinary brachiopods must haveused them to anchor themselves in soft,muddy seafloors. So these twosuccessful groups had conquered new

habitats and modes of life, and whoknows where the brachiopods mighthave gone but for the mass extinction.

The brachiopods were devastated bythe end-Permian event. At the level ofsuperfamilies, 16 out of 26 disappeared,which is bad, but not awful (it equates toa loss of 62%). However, at the level offamilies, 40 out of 55 died out (73%loss). It has been estimated that some95% of genera of brachiopods were hitby the extinction, and that equates toabout 99% of species. So all but a tinyhandful of this hugely diverse andabundant group bit the slime.

In contrast to this collapse of thebrachiopods, some of the molluscsweathered the end-Permian crisis much

better. There are three main groups ofmolluscs: the bivalves (‘two valves’),gastropods (‘stomach foot’) andcephalopods (‘head foot’). The originsof the last two of these names are ratherstartling. Gastropods do indeed have astomach in their foot, but their foot isactually almost their whole body.Technically, the soft slimy portion of asnail or whelk that creeps over theground is the foot, but obviously thewhole body of the animal, from eyestalks at the front to anus at the back, isenclosed in the foot. Cephalopodsinclude the octopus and the squid, aswell as the fossil ammonoids with theircoiled shells. The ‘head’ is the frontportion with its huge eyes and ring of

massive tentacles that haul food towardsthe mouth. The ‘foot’ consists of thetentacles and a siphon that can squirtwater or black ink for rapid jetpropulsion and confusion of an enemy.So, technically, the head and foot areclosely associated, and the ‘body’ of theammonoid, or of the octopus, is a bag-like structure containing the stomach andguts.

Most families of bivalves passedthrough the event relatively unscathed,with only three out of 40 disappearing.Bivalves in the Late Permian were rarerelements of the seabed faunas than werethe brachiopods, but they occupied arange of habitats, including living on thesurface of sandy and muddy seabeds,

and burrowing in the sediment. Theextinctions did not eliminate any of thesemodes of life, but species diversity was,as always, hit hard.

Gastropods include some ratherfiendish predators that prey on othershelled seabed animals such as bivalvesand arthropods. They attack them by anarray of mechanical and chemicalmeans, breaking and piercing through theshelly armour of their prey and suckingout the soft flesh. Others creep about onrocks, scraping off the green algae witha tough rasping radula, something like awoodworker’s mechanical sanding belt.Yet others hoover up organic detritusfrom the sea floor. Gastropods werediverse (i.e. lots of species) but rare

(i.e. not many individuals in anylocation) in the Late Permian, and theywere apparently hit harder than thebivalves, but they recovered well afterthe mass extinction. It has been estimatedthat perhaps 90% of gastropod generadied out, and this scales up to the loss ofthree out of 16 families. Most of thelosses were species that had limitedgeographic ranges or specialized diets:the more cosmopolitan forms survived,as did the generalist detritus-feeders.

The third major mollusc group, thecephalopods, were severely affected bythe mass extinction. The ammonoids, thefree-swimming forms with coiled shells,were nearly wiped out. Ammonoidshave always shown boom-and-bust

patterns of evolution. In good times theyevolved rapidly and diversified intohundreds of species worldwide. Then,when an environmental crisis camealong, they always died out. In the latestPermian Chinese sections, 20 out of 21genera, and 102 out of 103 species,disappeared at the very end of thePermian. The few species that creptthrough into the Triassic then radiatedrapidly, but the ammonoids virtuallydisappeared during the next massextinction at the end of the Triassic.After another huge and successfulrecovery from the brink, in the Jurassic,the ammonoids were finally anddefinitively killed off during the KTevent. Other cephalopods were much

less affected by the end-Permian crisis.The post-extinction world retained

only sorry remnants of the brachiopodsand bivalves, and they were fairlysimilar worldwide. An assemblageconsisting of the brachiopods Lingulaand Crurithyris, both leftovers from thePermian, is found everywhere in theEarly Triassic. Lingula in particular is awell-known disaster species, able tosurvive in all kinds of conditions,including low oxygen and low salinity.Indeed, Lingula is one of the mostastonishing ‘living fossils’, knownvirtually unchanged through most of thepast 500 million years. It might beexpected that such longevity would beassociated with some remarkable

features, but it is not. Lingula is abrachiopod shaped roughly like the nailon your little finger that lives buried inthe mud, held down by a long fleshy footand quietly pumping seawater through itsbody cavity so that it can extract foodparticles. The secret of its success ispresumably a wide environmentaltolerance, and the ability to survive onvery little food.

The post-extinction bivalves were asimilarly restricted group. In the EarlyTriassic, four or five genera occureverywhere. Although these supposedlycontain about 100 species, most belongto Claraia and Eumorphotis, both thin-shelled, scallop-like paper pectens (theyare called paper pectens because they

are both pectens, like modern scallops –the symbol of the Shell Oil Company –and they are also paper-thin) that wereattached by fine threads to irregularitiesthey found in the black anoxic EarlyTriassic muds. Cephalopods were rarein the Early Triassic, butmicrogastropods are suddenly, andbizarrely, hugely abundant worldwide.Then everything changed.

In the Mid Triassic, the uniform andcosmopolitan disaster species ofbrachiopods, bivalves andmicrogastropods were supplanted bybursts of evolution in all groups, but thebivalves and other molluscs radiatedexplosively. New species and lifemodes appeared, many of them with

distributions in individual ocean basins.By the end of the Triassic, the molluscshad recovered to Permian diversitylevels and had even begun to surpassthem. The brachiopods could notrespond fast enough it seems, and theynever recovered their Permian diversity.This episode has formed the basis of along-standing debate about the relativesuccess of the brachiopods versus thebivalves, which exemplifies a widerclash in our views of evolutionaryhistory.

The search for pattern

Does history follow rules or not? In

matters of human affairs, some historianssee a horrible inevitability in the riseand fall of nations. Countries andpeoples can be said to go through cyclesof rising fortunes, successful conquest,then decadence, followed by collapse.Marxist historians argue that there is apre-ordained plan, driven by socio-economic forces and the tension betweenthe ruling classes and the ruled. Otherswould use militaristic or martialmetaphors: nations rise and fall as aresult of their fighting power. Whenconfidence is high and the steel is sharp,that nation prevails.

In 1923, the Russian astronomer andarchaeologist A. L. Tchijevski publishedhis ‘Index of Mass Human Excitability’.

He argued that humans behave excitedlyand violently at regular intervals, ninetimes each century. Each cycle ofexcitability lasts for 11.1 years, and themaximum level of violence is associatedwith sunspot activity. In 1943, theAmerican psychologist Raymond P.Wheeler of the University of Kansasargued that civil wars follow a 170-yearcycle. Every third wave is supposedlymore dramatic than the others, givingphases of extreme violence every 510years. The cause in this model issupposedly periods of drought, whichapparently have happened every 170years. Other historians view suchnotions as utter nonsense, entirelywithout substance.

The search for pattern and meaningin history is an incredibly attractivepursuit. A quick search of the WorldWide Web reveals hundreds of sitesdevoted to showing how history followspatterns. The theories range from thecrackpot to the seemingly sane. Whyshould this search for pattern be sopopular? If you look hard enough youcan find cycles, patterns, regularities inany series of historical events. Suchclaims can only be justified if they canbe turned into predictions. So, themillennialist has to show by calculatingthe birth dates of Genghis Khan,Napoleon and Hitler, or whatever, thatthe rise of the next nationalist conquerorcan be predicted.

Pattern in history could all be in themind. Humans are tidy creatures wholike to file information away inmeaningful boxes. It is comforting to beable to catalogue all the kings andqueens of England, and classify them aseither good or bad, not something inbetween. Far easier to see patterns inhistory than to have to accept thealternative – a raw and uncontrolledseries of changes of immensecomplexity, and devoid of meaning. Thesame is true of evolution.

Evolutionary history was acomfortable phenomenon for the pre-evolutionists, and indeed for moderncreationists. God planned everything,and He imposed a clear pattern. Every

fossil had its place in the scheme ofthings, and the apparent coming andgoing of different groups of plants andanimals was all part of a parade towardsperfection. Whether these evolutionarychanges were seen as unidirectional(time’s arrow) or cyclical (CharlesLyell’s view), there was a goal and ameasure of predictability.

Early theories of evolution, such asthat of Jean Baptiste Lamarck, were noless goal-directed – what thephilosophers call ‘teleological’. Assimple organisms evolved into morecomplex organisms, the next stage belowmoved up a step in the great chain ofbeing. Darwin forced people to rejectsuch comforting views. If evolution

happened by natural selection, it couldnot be predictable. But even Darwincould not entirely reject some of thecomforting pattern of Nature. Heretained the pre-evolutionary concept ofplenitude, literally ‘fulness’. Plenitudeencompassed the idea that everythingthat could be done in Nature was beingdone: the Earth was full of species ofplants and animals, and that number waspretty well fixed. In a famous analogy,Darwin wrote:

Nature may be compared to a surface coveredwith ten-thousand sharp wedges… representingdifferent species, all packed closely togetherand driven in by incessant blows,… sometimesa wedge of one form and sometimes anotherbeing struck; the one driven deeply in forcing

out others; with the jar and shock oftentransmitted very far to other wedges in manylines of direction.3

In another analogy, Darwin comparedlife at any time to a flotilla of applesfloating on the surface of a huge barrelfull of water. If a new apple is to insertitself in among the floating mass, it hasto push a pre-existing apple either up ordown.

So, whether wedges or apples, inDarwin’s view species could come andgo through the course of evolution, but anew species could not arise and find itsplace without first displacing a pre-existing species. This seems to makesense – or does it?

The race is not (always) tothe swift

In fact there is no evidence for plenitude.The history of life through long spans oftime shows that species diversity hasrisen time after time. New spheres forlife have opened up the opportunity forbursts of diversification: for example,when life first moved on to land, whenthe first forests appeared, when the firstinsects took to the air, when the firstcoral reefs arose in the sea, whenvertebrates evolved warm-bloodedness,and so on. A study of plants and animalsthat have become extinct in the pastmillion years shows plenty of empty

niches. What creature today is doingwhat the mastodons, mammoths andwoolly rhinos of North America andEurope did until 10,000 years ago? Whatabout the giant ground sloth of North andSouth America, which fed on caper andmustard plants, and which disappeared11,000 years ago? It would bepresumptuous in the extreme to assumethat all potential for the expansion of lifeinto new habitats and niches has come toan end just at the moment when wehappen to be here and worrying aboutour place in Nature. Darwin’s barrel ofapples has more to do with the limits ofour imagination than with reality.

So, what about the end-Permiancrisis and the brachiopods and bivalves?

This was for a long time a key exampleof plenitude in action. The brachiopodshad to go in the end because the superiorbivalves ousted them step-by-stepthrough the Palaeozoic. Comparisons ofmodern brachiopods and bivalvesseemed to support this view:brachiopods have a limited array ofmodes of life, while bivalves live in awide range of habitats, swimming,creeping and burrowing on sea beds, onshores, and in rivers and lakes aroundthe world. Brachiopods eat very littleand don’t do very much, while bivalveseat much more and (some at least) arevery active. Here was a classic exampleof competition on the large scale. Notcompetition between two individuals for

a mate, or between two species for foodor space, but between brachiopods as awhole and bivalves as a whole. Therewas some sort of justice about the rise ofthe mighty clam, and the demise of themiserable brachiopod.

In 1980, Stephen Jay Gould and JackCalloway challenged this comfortablevision.4 First, they asked, how could thearms race between brachiopods andbivalves have sustained itself in thePalaeozoic seafloor ooze for 300million years? When ecologists studycompetition between modern organisms,they expect to measure selectiveadvantages on time scales of years ortens of years at most. Over hundreds ofmillions of years, the selective

advantages would be so minute as todisappear in the statistical variation ofnormal populations. Gould andCalloway then made counts of thenumbers of genera of brachiopods andbivalves through the Palaeozoic andpost-Palaeozoic. It turned out that theexpected steady rise of bivalves andsteady decline of brachiopods had nothappened (Fig. 30). Both groupsremained at pretty well constantdiversity through the Palaeozoic, thebrachiopods with 150–200 genera, andthe bivalves with 50–100. Then camethe end-Permian crisis. Both groupswere hit hard, as we have seen, and onlyabout 50 genera of each survivedthrough into the Triassic. What was

critical was the post-extinctionrecovery.

In the Triassic, the bivalves made itand the brachiopods did not. Bivalvediversity climbed steadily from 50genera to over 100 by the end of theTriassic, and continued rising to morethan 400, while brachiopod diversityremained firmly at 50 or so. So,concluded Gould and Calloway,brachiopods and bivalves were notinvolved in continuous hinge-and-lophophore struggle throughout thePalaeozoic. In fact, their modes of lifewere often quite different, andcompetition between an averagebrachiopod and an average bivalvewould be as ludicrous as suggesting that

rabbits and snails compete, just becausewe find them together and they both eatplants. In Gould and Calloway’s words,brachiopods and bivalves were ‘shipsthat passed in the night’, groups thatlived side-by-side but did not seriouslyinteract. The competitionists wereincensed by Gould and Calloway’sanalysis, but the facts have to lead to therejection of the comfortable competitiontheory.

30 A classic example of supposedcompetitive replacement. Thesteady rise of the bivalves throughthe past 500 million years wasonce thought to have caused thedemise of most brachiopod groups.However, it seems that thebrachiopods and bivalves neverinteracted in a major way. It wasthe end-Permian crisis that hit bothgroups hard, and only the bivalveswere able to recover.

‘Comfortable competition’ may seeman odd juxtaposition of words. But, inevolutionary biology there has been apervasive view that competition (that is,the interaction between any pair oforganisms in which one gains a benefitat the expense of the other) guides the

large patterns of evolution. Such aviewpoint leads to explanations oflarger-scale patterns, such as the riseand fall of dynasties of plants andanimals, migrations, the sequence ofgroups of organisms through time, andextinctions, by competition. Themathematics of competition arerelatively tractable. So competition iscomfortable for many biologists andtheorists.

Competition can also be seen as partof the desire for predictability andpattern. Just like Tchijevski and hiscycles of human excitability, we find itvery hard to let go of the reassurance ofnumbers and meaning, whether in thehistory of human societies or in the

history of life. Far bleaker to have toaccept that nation states come and go,animal and plant groups appear and dieout, and there is no pattern, no innatesystem guiding their relative fates.

Perhaps the author of Ecclesiastes(9:11) got it right long ago:

I see this too under the sun: the race does notgo to the swift, nor the battle to the strong;there is no bread for the wise, wealth for theintelligent, nor favour for the learned; all aresubject to time and mischance. Man does notknow his hour; like fish caught in thetreacherous net, like birds taken in the snare, sois man overtaken by misfortune suddenlyfalling on him.

Trilobites: did they ordidn’t they?

All classic accounts of marine life of thePalaeozoic say that the trilobites wereone of the most striking groups to havedied out at the end of the Permian. Other,more detailed, overviews say the exactopposite: the trilobites were notinvolved. Both views are right. But, howcan such contradictory statements bemade?

Trilobites5 had indeed dominated thesea-floor since Cambrian times. Withtheir three-lobed bodies, consisting of amidline body portion and two lateralareas that bear the eyes at the front and

cover the legs and gills further back,trilobites were the most active, andseemingly most complex, animals of thePalaeozoic. They belong to the PhylumArthropoda, a huge group of animals thatincludes insects and spiders, as well asmarine forms such as crabs and lobsters.Arthropod means ‘jointed limb’, an aptname for animals which have a toughouter skeleton like a suit of armour withflexible joints all along the legs.Trilobites ranged in size from themicroscopic to over 30 centimetres long.Many of them had complex eyes and,presumably, excellent vision, as well aselectrical sense organs. They livedlargely on the sea bed, ploughing throughthe surface mud feeding on detritus.

Many of them could swim, and somemay have hunted actively.

Trilobites were common and diverseespecially in the Lower Palaeozoic, andthey survived through to the Permian, buthad disappeared before the Triassic. Itis true, then, to say that they becameextinct in the Permian, but the class wasalready much reduced before thebeginning of the Permian. Only threesmall families of trilobites are knownfrom the Permian. One or two speciesare known from the latest Permiansections in China, and these indeeddisappeared at the Permo-Triassicboundary. The final disappearance of thetrilobites was a fairly feeble little blipin their history, and it followed a long-

term decline. So, although the rule of thetrilobites finished at the end of thePermian, closer analysis shows that thegroup was already very much on the wayout.

While other arthropod groups in thesea, including ancestors of modern crabsand lobsters, were apparently littleaffected, an unusual arthropod group, theostracodes, were hit relatively hard.Ostracodes are generally tiny, amillimetre or less across, and they looklike tiny animated beans. Their bodiesare enclosed in two tightly fitting valves,and they swim and feed by opening thevalves and extending feathery little legsand gills out into the water. The only factabout ostracodes that my students ever

remember is that the males have a penisthat is often longer than their bodies.Shallow-water ostracodes suffered somemajor extinctions, while the deeper-living, more cosmopolitan groups wereapparently little affected by the end-Permian crisis.

Fishes: extinction orLazarus taxa?

As with the arthropods, the fossil recordof fishes has been interpreted in twoways. Some commentators say that thefishes simply swam unconcernedlythrough whatever was happening at the

end of the Permian, just as they wereseemingly unaffected by most other massextinctions. Others claim that theysuffered extinctions, like other groups.Who is right?

Certainly the supporters of fishsurvival can make a case. They note thatall the major fish groups from thePermian are found in the Triassic. Buttheir opponents argue that the sharksshow major changes in size. The diversearray of medium- and large-sized sharksof the Late Permian was reduced to afauna of only small sharks in the EarlyTriassic (see Figs 28 and 29). Whetherit was only small forms that managed tosurvive, or whether the survivors,variable in size, then became dwarfed

because of some evolutionary pressureis uncertain. Also, among the bony fishes– the group that today includeseverything from haddock to goldfish, andtuna to seahorse – two out of eightfamilies disappeared. But did they reallydisappear?

The extinction apologists point to thepossibility of Lazarus taxa. Lazarus taxa,are species or genera or families thatseem to disappear, and thenmiraculously reappear. Well, not‘miraculously’. Their temporary absencemay just be a failure of preservation.The term comes from the famous biblicalstory of Lazarus (St John, 11: 17-44):

On arriving, Jesus found that Lazarus had beenin the tomb for four days already.… The dead

man came out, his feet and hands bound withbands of stuff and a cloth round his face. Jesussaid to them, ‘Unbind him, let him go free’.

In a palaeontological example, it is muchmore likely that a group that apparentlydisappears and then reappears later hasactually been there all the time, and theapparent absence, or death, merelyrepresents a gap in the fossil record.

So, in the case of the end-Permianfishes, there was an apparentdiversification of bony fishes in theEarly Triassic, when nine new familiesappeared, or ‘appeared to appear’.Perhaps an extinction followed soonafter by a diversification is really tellingus that there was simply a gap in thefossil record there. The animals lived on

through the intervening time, but theirskeletons were not preserved. Actually,this explanation is not good enough, andthe families of fishes that apparentlydisappeared at the end of the Permiandid not reappear anywhere. The newfamilies in the Early Triassic are indeednew families. And there does truly seemto have been a phase of dwarfing of theEarly Triassic fishes, which oddcircumstance may be telling ussomething about the environmentalcrisis.

Survivors and victims

So, in the end, despite much debate

about individual groups, current studiessuggest that there were major extinctionsat the end of the Permian among virtuallyall marine creatures. Groups, like thefishes, that were apparently littleaffected, turn out, on closer study, tohave suffered just as much as the others.Older statements that some groupssurvived unscathed often turn out to havebeen based on incomplete and perhapsrather general examinations of the data.The more detailed, site-by-site, bed-by-bed, studies that are now being doneshow us more exactly what was goingon.

A comparison of the survivors andthe victims can provide clues about thecrisis. One might hope to find evidence

that, say, carnivorous animals wereharder hit, or perhaps tropical dwellers,or large animals, or specialists, orwhatever. Actually, it is very hard tofind evidence for selectivity of this sort.In general, of course, geographicallywidespread species survive better thanthose restricted to small areas. Inaddition, those that have wide tolerance– animals that can survive in a widerange of temperatures or on a variety ofdifferent food types – might also beamong the survivors. Any plant oranimal that is adapted to a very narrowhabitat and that feeds on a small range offoods is likely to be vulnerable duringany crisis.

In a detailed survey of the victims

and survivors of the end-Permian crisisin the sea, Tony Hallam and PaulWignall6 argue that the survivors doshare one key feature – an ability to livein low-oxygen conditions. This is true ofthe surviving foraminifera, brachiopods(at least Lingula and Crurithyris),bivalves (the paper pectens Claraia andEumorphotis) and ostracods.

Dwarfing was another feature of thesurvivors, most notably the gastropodsand fishes. In the Late Permian, thegastropods and the fishes display abroad range of sizes. The same is truefor the Mid Triassic. But in the EarlyTriassic, a new rock type is foundworldwide: microgastropod grainstone.This is a limestone made up solidly of

billions of millimetre-sized shells ofgastropods. Perhaps less dramatically,sharks and bony fishes of the EarlyTriassic were all a size or two smallerthan either before or after that period.

Survival in low-oxygen conditionsand dwarfing are both perhaps responsesto low productivity seas. The crisiswiped out so many organisms – 90% ormore of all species had disappeared –that established food webs and foodchains broke down. Normal processes oftransmitting energy and carbon from theplankton up through ever-largerpredators were destroyed. Perhaps thosesmall and undemanding organisms thatdid not need much food, or much oxygen,were the only ones that could eke out a

living in the post-disaster conditions.

Rarefaction – rarefiction?

The approximate figures for the scale ofextinctions during the end-Permian crisiswere given at the beginning of thechapter. Some 50% of marine familiesdied out, and this scales to 90 to 96% ofspecies. The higher estimate of 96% wasmade by the University of Chicagopalaeobiologist David Raup7 in 1979.He used a mathematical method calledrarefaction, which is a way of estimatinghow low-level effects, for example theloss of individual species, feed up to

higher-level phenomena, such as theextinction of families. We saw thisscaling effect in considering how theextinction of a relatively smallproportion of brachiopod superfamiliesimplied the extinction of a very highproportion of genera and species.

Critics suggested that Raup hadperhaps been a little over-enthusiastic.Toni Hoffman, a Polish palaeobiologist,suggested a more conservative 75%species loss, arguing that the rarefactionapproach could not be regarded asreliable. Mike McKinney, another critic,accepted the validity of the rarefactionapproach, but he came up with a revisedfigure of 90% species loss. This revisedfigure followed from a detailed

reconsideration of the rarefactionmethods. He found that Raup hadoverestimated the species extinctionfigure since he had assumed a randomdistribution of species extinctions acrossthe evolutionary tree. But this, arguedMcKinney, is not correct.

McKinney noted from his studies ofsea urchin evolution that the patterns ofextinction of species within genera werenot uniform. In other words, certaingenera, each containing several species,were more likely to go extinct thanothers. If species loss is not randomlydistributed, then some genera will havea higher chance of going extinct thanothers. Equally, others will survivebetter. Overall, this non-randomness

means that a fixed level of family orgeneric extinction can be founded on alower species extinction rate than anentirely random distribution.

So the best estimate for species lossat the end of the Permian is 90%. This isless dramatic perhaps than a loss of 96%of species, but it is still by far the largestmass extinction of all time, and certainlythe closest that life has ever come tocomplete annihilation.

Palaeontologists have alwaysaccepted that the end of the Permianmarked a major crisis for life in the sea.But their views about the extinctions onland have been much more mixed.Indeed, some claimed that nothing muchhappened at all. New research has

overthrown that idea, as we shall see inthe next chapter.

9

A TALE OF TWOCONTINENTS

There seems to be little doubt about thescale of the end-Permian crisis in thesea. Ever since the work by JohnPhillips in the 1840s, palaeontologistshave accepted that this was a time ofmajor turnover in the history of marinelife. But what of the land? The story has

been very different. As we saw earlier(Chapter 4), many, perhaps most,palaeontologists once happily deniedthat there had been any more than ahiccup in the history of the reptiles andamphibians.

The same was also true for expertson fossil plants and fossil insects; theysaw their favourite fossil groups inPermian rocks and in Triassic rocks, anddid not recognize any major differences.Closer study now shows that the crisison land was just as severe as in the sea.Why this apparent failure to recognizethe truth?

Part of the problem has always beenthe notoriously patchy terrestrial fossilrecord. Marine palaeontologists rejoice

in great thicknesses of rock, sometimeshundreds of metres, even kilometres, oflimestones and mudstones laid downyear-by-year on ancient continentalshelves, and stuffed with fossils. Thesame is rarely true for the continentalfossil record. The term ‘continental’ isused as a proper equivalent to marine,meaning the combination of terrestrial(ancient soils, screes, landslides),lacustrine (lake) and fluvial (river)deposits. Even if the plants and animalsof interest lived completely on land,their fossils are commonly found insediments deposited by ancient rivers orlakes, since most of the land is notassociated with deposition. In fact quitethe opposite. Mountains and hills are

sites of erosion, so rocks and fossils arerarely found there. Lowland soils arerare also, since most of them have atransient existence, deposited, rained on,washed away, redeposited and finallycarried into the sea. Rivers pick upsediment and inexorably wash itdownhill until it finally enters the sea atan estuary, when it may be transportedfurther across the continental shelf byslumps and turbidity flows until it mayfinally enter the deep oceanic abyss.

There might seem, on reflection, tobe little hope of finding any continentalfossils. Not so, fortunately. There aresome settings on land where sedimentscan accumulate in considerablethicknesses. Most notable are ancient

lake systems in subsiding basins. Forexample, rift valleys are sites where acontinent unzips. A good modernexample is the Great Rift Valley runninghalfway down the eastern half of Africa,from Ethiopia to Mozambique. Thecontinental plate carrying Africa is in theprocess of pulling apart, and a greatocean may some day flood into the split.Over the past 20 million years, however,the rift has been filling with sediment, toa depth of kilometres in places.Palaeontologists can follow theevolution of their terrestrial fossilgroups of interest, whether they besnails, grasses or humans, through short-term, even annual, layers.

The great Karoo Basin

In the Late Permian, southern Africa laynear the south pole. The base rock of thatpart of the continent had been formedfrom molten rocks thousands andhundreds of millions of years earlier.During the Permian, Africa was part ofthe supercontinent Gondwana. This wasa single, fused landmass with nodivisions, and to draw the moderncontinents over it is somewhatmisleading. None the less, it isimpossible to visualize the nature ofGondwana without making a mentaljigsaw reconstruction using the familiarshapes of the modern continents. SouthAmerica snuggled up close against the

western margin of Africa, with theeastern bulge of Brazil fitting neatly intothe Congo Basin. Land extended from theeastern margins of Africa throughMadagascar, now an island, and acrossto India. The triangular shape of Indiatoday relates to its ancient Gondwananposition, when it fitted, like a wedge,into the northern margin, squeezed upagainst the eastern coast of Africa andMadagascar. To the south, Antarcticawrapped itself around the southern tipsof South America, Africa and India. Theeastern margin of Gondwana was madefrom Australia, rotated so that its southcoast faced west and fitting neatlyagainst the eastern side of India andAntarctica.

A great basin, the Karoo, forms theheart of South Africa today. The scrub-covered land of this great desert-likebowl offers poor returns to the farmers,but it is rich territory for geologists andpalaeontologists. In the Permian it wasringed by mountains to the south, lyingon what is now Antarctica. The basin is1500 kilometres across, and up to 10kilometres of sediment accumulatedfrom the Early Permian to the EarlyJurassic. During this time, the wholebasin was subsiding, partly as a result ofancient fault systems that remainedactive and allowed the underlyingbasement rocks to sink ever deeper, andpartly from the accumulating weight ofsediment, which added to the downward

pressure. Sediment thicknesses aregreatest in the south of the Karoo Basin,good evidence that that was the maindirection of supply. In geology, theproximality principle applies: the closeryou are to the source of some rocks, thethicker they will be, whether they arevolcanic lavas or river-depositedsediments. So, even though themountains that supplied rocks, sand andmud to the Permo-Triassic rivers ofSouth Africa have long gone, there is nodoubt that they existed, and that they layto the south.

The sediments of the Karoo Basinspan the Permo-Triassic boundary, andthey are full of plants, insects andreptiles. Ideal territory to study the

nature of the end-Permian event on land.

Professor Owen and theScottish engineer

The Karoo reptiles first came to publicnotice in 1845. In 1844, Henry De laBeche, ‘foreign secretary’ of theGeological Society of London (andcreator of Professor Ichthyosaurus; seeFig. 8) had received a long letter fromGrahamstown in South Africa fromAndrew Geddes Bain (1797–1864), aScottish engineer who was employed inmilitary road building in what was thenknown as Cape Colony.1

Bain had been an enthusiasticgeologist ever since he had read CharlesLyell’s Principles of Geology in the1830s, and he never ventured out ofdoors without his hammer and collectingbag. In the course of his work in thesouthern parts of the Karoo Basin, Bainhad stumbled across dozens of reptilespecimens, the first in 1838. He hadextracted them from the rocks with careand had made a collection of manydifferent species, some as small as amouse and others as large as arhinoceros. In his letter to De la Beche,Bain described the reptile finds, andalso reported on the geology of the rocksin which they had been found. Thefossils arrived at the same time as the

letter, shipped in several packing casesfrom Grahamstown.

The committee of the GeologicalSociety immediately sent Bain theWollaston Fund, a sum of 20 guineas(£21), as a reward for his work, andthey encouraged him to continue tocollect. Bain’s fossils were handed toRichard Owen for description. Shortlyafter, a paper about the geology of theKaroo Basin, extracted from Bain’sletter, was read to the GeologicalSociety, followed by Owen’sconsideration of the reptile specimens.2

Owen recognized, as Bain had, thatmost of the new South Africanspecimens belonged to a single groupwhich Owen termed the dicynodonts.

These creatures had a short, high skull,and apparently a horn-covered turtle-like jaw (Fig. 31). They typically hadonly two teeth (Bain had called thesebeasts ‘bidentals’), and Owen’s termDicynodon meant ‘two-dog-tooth’, inreference to the fact that the pair of teethwere the canines, the sharp teeth that areparticularly long in flesh-eatingmammals today and in humans are morepointed than the neighbouring incisors infront and cheek teeth behind. Thedicynodonts had rather massive bodies,with stumpy legs, and a short, one mightsay disappointing, tail.

Owen’s considerations of the Karooreptile specimens came before hisdescription of the isolated reptile bones

from Russia (Chapter 1), in March 1845.He looked at Bain’s bones in late 1844,and read his first account to theGeological Society of London in January1845. Little did he know it, but Owenhad in his hands, from Russia and fromSouth Africa, at the same time, evidencefor a major new group of reptiles –evidence that could have revolutionizedpalaeontological understanding. But theRussian fossils were too incomplete forhim to spot the similarities, and the linkwas made only a few decades later.

Bain sent back more collections offossil reptiles to London, and Owen wasable to publish further descriptions.Indeed, Owen made sure that AndrewBain, and later his son Thomas, were

well remunerated by British governmentsources. For example, in 1877, ThomasBain was sent £200 to carry out a large-scale fossil-collecting expedition in theKaroo Basin by ox cart. The tripcoincided with a severe drought, andBain was hard pressed to find enoughwater to keep his men alive.Nevertheless, he reported to Owen thathe had come back with ‘almost 280heads’ of Permo-Triassic reptiles,which were all sent to London for Owento describe and name.

31 The skull of the plant-eatingsynapsid Dicynodon.

Later in the nineteenth century, theyounger British palaeontologists ThomasHenry Huxley and Harry Govier Seeley(1839–1909) also began to describe andname new South African specimens.Seeley went to South Africa in 1889 tolook at specimens in the museums therefor himself, but he found that the bestones had already been sent to the British

Museum in London. Nevertheless,Seeley did collect some furtherspecimens, and brought them back toLondon.

Wing collars and addictionto fossils

Everything changed in the twentiethcentury. It became less common forcolonial fossils to be sent back toEurope in general, whether Britain,France or Germany, and South Africagained independence from Britain in1902, following the Boer Wars. Inaddition, the arrival there of an energetic

palaeontologist by the name of RobertBroom (1866–1951), a Scottish-bornphysician, meant that there was nolonger any need to seek advice frompalaeontologists based in Europe. Aftertraining as a doctor in Glasgow, Broom3

went first to Australia to build his careeras a country doctor. In his spare time, hestudied fossil and modern marsupials inhis adopted country. Then, on a visit toLondon in 1896, he saw the SouthAfrican fossil reptiles in the BritishMuseum, and he was hooked. Hedecided not to return to Australia,emigrating instead to South Africa wherehe became a country doctor again, butturned his attention to the burgeoningcollections of Permo-Triassic reptiles.

He was always properly dressed, andconducted his field work in a formalcoat and hat, and a stiff wing collar, aswas the Edwardian fashion, even in theblistering African sun. However, it issaid that sometimes it became too hoteven for him, and he would strip nakedto cool down.

By 1900, collectors in South Africawere so overwhelmed with fossilspecimens that they concentrated simplyon picking up skulls; the skeletons wereleft to erode away. Broom was soonsending descriptions and reviews of theSouth African Permo-Triassic reptiles toscientific journals in London, andincreasingly to those of the scientificsocieties in his adopted country. In the

early 1900s, he managed to write a merefour or five scientific papers each year,but this soon rose to a steady twenty orso. By the time of his death in 1951 hehad written a total of over 400 scientificpapers, as well as several books.

Clearly, Broom worked fast and hehad a prodigious memory – he is said tohave been able to remember everythinghe had ever written and seen, and indeedeverything he had ever read. Also, heapproached his God-given duty to find,name and describe new fossils with zealand immense speed. Typically, he wouldwrite a description and make freehanddrawings of his new finds in a matter ofdays, and send off the manuscript to thepress within a week. Despite this haste,

not uncommon at the time but quite acontrast with the slow and painstakingwork that is more normal today,Broom’s observations were generallyaccurate. His one failing, and again thiscan only really be said in retrospectsince he was following the lateVictorian norms he had learnt from hispredecessors, was that he named far toomany new species and genera. He saw anew form in every half-decent fossil,and named over 200 new species ofPermo-Triassic reptiles.

Precursors of the mammals

The majority of reptiles named by

Owen, and later by Broom, from theKaroo Basin were mammal-like reptiles.At first, however, Owen had beenunclear just what they were. He toyedwith the idea that Dicynodon was arelative of turtles, based on its generallytoothless jaws, pointed beak and solidskull bones. Then he thought it might besomehow related to modern lizards,although his logic was somewhat arcanein that case. But in the end, he realizedthat these reptiles were related in someway to mammals. In his words, theseKaroo predators ‘made a most singularand suggestive approach to themammalian class.’

The most obvious clue is in the teeth.Instead of the typical reptilian

arrangement of essentially identical teethfrom fore to aft, the jaws of thesynapsids, as the group is termed (seeChapter 1), show a clear flesh-eating,mammal-like complement of teeth: smallspiky incisor teeth at the front, then along curved canine tooth and a numberof cheek teeth behind (Fig. 32). Othermammal-like features include a dog-likeface, a single lower opening in the skullbehind the eye socket, various featuresof the skeleton, and even hair. Synapsidhair is not preserved in any LatePermian fossils, but some earlycynodonts have masses of small nerveopenings in the snout region, whichimply that they had whiskers,presumably for sensory purposes – and

if they had whiskers, they also had bodyhair.

The early synapsids are often calledpelycosaurs, and their most famousrepresentative is the iconic Dimetrodon.Dimetrodon has a smiling face, sharpflesh-eating teeth and a vast sail on itsback. Commonly mis-classified as adinosaur, Dimetrodon graces everychild’s book of prehistoric beasts orcollection of plastic models. The sailalong the ridge of its back is made fromskin stretched over the enormouslyextended spines attached to itsbackbone. Whether the sail was used forthermoregulation or not (and that is whatevery book says), most pelycosaurslacked a sail. By Mid Permian times,

perhaps 260 million years ago, thepelycosaurs had given way to the newbreed of synapsids, termed thetherapsids (‘beast arches’). The ancientGreek ther, therion means ‘beast’, andis interpreted as equivalent to mammalin more correct modern terminology.

The Karoo therapsids include abasal, and varied, group of meat-eatersand plant-eaters, the dinocephalians,some of them quite large. The plant-eating Moschops in particular was anextraordinary animal, equipped withmassive shoulders and neck and tinylimbs and head. Moschops also had amassively thickened skull roof, and ithas been suggested that it engaged instentorian head-butting contests in

seeking mates. The meat-eatingdinocephalians were more lithe andsuperficially dog-like.

The commonest herbivores were thedicynodonts, with either no teeth at all ormerely a pair of tusks (see Fig. 31).They sliced vegetation with their sharp,horn-lined beaks, and ground it with acircular jaw motion. The dicynodontsand large plant-eating dinocephalianswere preyed upon by the first sabre-tooths, the gorgonopsians.Gorgonopsians were powerful sleekanimals with massive tusks. They coulddrop the lower jaw clear of the fangs,and then leap at a herbivore with theteeth exposed and pierce their thickskins.

32 The skull of the flesh-eatingsynapsid Thrinaxodon from theEarly Triassic of South Africa.

The light dawns …

Unknown to Owen, synapsids hadalready been independently describedfrom the Late Permian Copper Sandstonebeds of the Ural Mountains in Russia bya number of palaeontologists there – F.Wangenheim von Qualen, S. S. Kutorga,J. G. F. Fischer von Waldheim and E. I.

von Eichwald, all of whom wereintroduced in Chapter 1. Owen is notreally to blame, however, for not makingthe link at that time. These impressivelyGermanic gentlemen published ratherlimited accounts, some of them insomewhat obscure journals in StPetersburg and Moscow. Moreimportantly, in the 1840s, neither they,nor Owen, had any concept of synapsids.

Enlightenment was rather slow tofollow. Owen saw all the new SouthAfrican materials first-hand, but he wasnever able to visit Russia to confirm thatthe Permian reptiles there wereessentially the same as those from theKaroo. The first light was shed in 1866when Hermann von Meyer (1801–69),

the senior German palaeontologist of theday, published a large monograph on theamphibians and reptiles from the Urals,based on a collection of von Qualen’sfossils that had been sent to Germany.4

The connection was completed ten yearslater when the British Museum acquiredits own collection of Copper Sandstonereptiles and Owen was able to comparethem with the more complete Karoospecimens. He then concluded that therewas a firm connection between thereptiles of the Russian CopperSandstone and those from South Africa.

Another English researcher, WilliamHarper Twelvetrees, published severalarticles in 1880 to 1882 on vertebrateremains and other fossils from the

Kargala mines in the Urals, where hehad worked as a mining engineer for anumber of years. Twelvetrees emigratedto Australia in 1891 and was employedas director of the new TasmanianGeological Survey which was launchedin 1899. But it was Harry Seeley whowas probably the first person to visitboth South Africa and Russia, and to seethe Permo-Triassic rocks, and theirenclosed reptiles, himself. Havingvisited the Karoo Basin in 1889, hepublished5 an account of the collectionsof Russian Permian reptiles in the SaintPetersburg Mining Institute and those ofKazan University.

Copper mining in the Urals hadbecome uneconomic by the mid-

nineteenth century, largely because ofcompetition from new and more efficiententerprises in Africa and Australia. Thelast copper mine in the South Uralsclosed by 1900. New discoveries ofvertebrates from the Copper Sandstonesalmost completely stopped, except foroccasional finds – which can stilloccasionally be made – of scraps ofbones in the spoil heaps of the oldmines.

A time-scale of the crisis

With all these reptile specimens fromboth South Africa and Russia, it shouldperhaps have been straightforward to

follow through time exactly what hadhappened at the end of the Permianperiod. Surely the palaeontologistscould readily track, bed-by-bed, thecollapse of the Late Permian faunas, andthe rise of new forms after thecataclysm?

In reality, of course, things are not atall so simple. A great deal ofgroundwork had to be done during thetwentieth century before geologistscould begin to disentangle the exactnature of the huge catastrophe that hadtaken place on land. The puzzle was, hadall these creatures been living at thesame time, or did they constitute asequence of faunas that in fact spannedmillions of years? Bain, Broom and

others had begun to divide up thesequence of rocks in the Karoo into‘zones’, each representing a unit of time,associated with and named afterparticular reptiles. But there wereproblems.

The synapsid skulls and skeletonshad been extracted from sites all overthe huge area of the Karoo Basin.Precise levels in the rock sequencecould not be determined clearly, mainlybecause the Permo-Triassic successionlooked very similar throughout – redsandstones and mudstones. There was nosingle, simple vertical reference sectionin which the horizons could beidentified. So two skulls found 1000kilometres apart could well be from the

same species, and from rocks of thesame age. Equally, two finds from thesame farm could turn out to be utterlydifferent, and separated in time by 20million years.

A second major problem was that noone in 1900 had a clear view of thetime-scale involved in the Permo-Triassic, nor of how to match thecontinental sequences from South Africa,Russia, or indeed from England orGermany, with marine successionsknown from the Alps.

The South African Karoo sequencewas known to consist of a long run ofcontinental sediments, ranging in agefrom the Carboniferous to the Jurassic.At the base of the sequence is the Dwyka

Tillite, a deposit made from the ground-up rocks deposited beneath glaciers.Here was evidence for the huge south-polar glaciation of the LateCarboniferous and Early Permian. TheDwyka Group also includes, above thetillite, Early Permian beds with fossilplants. Above the Dwyka Group (700metres thick) comes the Ecca Group(2000 metres thick), largely MidPermian in age, and this is followed bythe Beaufort Group (up to 4500 metresthick), ranging in age from Late Permianto Mid Triassic. The famous Karoofossil reptiles virtually all come fromthe Beaufort Group. Above this lies theStormberg Group (500 metres thick), aLate Triassic to Early Jurassic

succession of continental red-bedsediments. In some horizons, the remainsof early dinosaurs have been found, inthe form of both skeletons and footprints.The sequence is capped by theDrakensberg volcanics, basalt lavas thatpoured out in the Early Jurassic asGondwana began to rift into itsconstituent modern continents. In 1900,geological understanding of the KarooPermo-Triassic rock succession waswell ahead of that in Russia. But by1950, the situation in Russia had greatlyadvanced, largely due to the work of oneman.

The father of the Russian

Permo-Triassic

Ivan Antonovich Efremov (1907–72)was the right man at the right time (Fig.33).6 With a brilliant synthetic mind, hewas able to draw together the seeminglyunconnected records of Permo-Triassicvertebrates from the vast Russianterritories, and cajole them into ameaningful stratigraphic scheme. Notonly this, but Efremov is credited withthe invention of a new branch ofpalaeontology, which in 1940 he calledtaphonomy, the study of the preservationof fossils. He was also a writer ofscience fiction novels, which werehighly popular in Russia, both for theirimaginative story lines and for the

mildly pornographic images of womenwhich he used as illustrations.

Efremov’s scientific activity beganin 1925, when he was 18 years old. Hewas employed as a fossil preparator inthe Mining Museum in Leningrad and hebegan straight away to lead fossil-hunting expeditions south to the CaspianSea, east to the classic Urals Permo-Triassic and northeast to the DvinaRiver. The North Dvina River is a hugewaterway that begins high in the UralMountains and runs for 1000 kilometres,heading first southwards, then east andnortheast to the icy port of Arkhangel’sk(often called Archangel in English),where it enters the Arctic Ocean.Efremov took over the major expeditions

that had been begun on the North Dvinaby V. P. Amalitskii (1860–1917),professor of geology at Warsaw inPoland. Working against considerableobstacles, and with very little finance atfirst, Amalitskii had proved the richnessof the Late Permian beds along the NorthDvina. Over years of return visits, heand his wife extracted hundreds ofskeletons of a whole new array ofamphibians and reptiles, some of themunique to Russia, such as the amphibiansDvinosaurus and Kotlassia, and thesabre-toothed gorgonopsianInostrancevia (see Fig. 35). Efremovcollected more amphibians and reptilesalong the Dvina River, and made surethat Amalitskii’s unfinished work was

completed.

33 Portrait of Ivan AntonovichEfremov, who worked out thestratigraphic scheme for theRussian Permo-Triassicvertebrates.

Nearer to home, Efremov excavatedsites north of Moscow, where he found

some of the first Early Triassicvertebrates in Russia, including skulls ofthe fish-eating amphibian Benthosuchus.He also reinvestigated the copper minesin the South Urals, crawling about insome of the old shafts in the hope ofmaking new finds. While he wasdisappointed in this aim, he was able toprovide an overview of what had beenfound, and the conditions of theirpreservation. More promising were thePermo-Triassic red beds that lay aroundthe South Urals. Murchison had visitedthese in 1841 and had collected fromthem the reptile bones he passed toOwen. Very little had been done since.But Efremov and his colleagues began toidentify numerous localities on both

sides of the Ural and Sakmara rivers.They noticed how certain species tendedto occur together regularly – that thereappeared in fact to be a succession ofquite distinctive faunas.

Efremov recognized seven zones:

Zones I, II: Dinocephalian (LatePermian)Zone III, IV: Small reptile andpareiasaurian (Late Permian)Zone V: ‘Neorachitome’ (EarlyTriassic)Zone VI: Capitosaurus (EarlyTriassic)Zone VII: Mastodonsaurus (Midto Late Triassic)

The division between Zones IV and Vmarks the line of the end-Permian crisis.Each zone is defined by a characteristicreptile or amphibian, part of a typicalassemblage. The Late Permian Zones I–IV are named for particular reptiles,while the Triassic Zones V–VII arefounded on aquatic amphibian species.These zones seemed to work throughoutRussia, from Moscow to the Urals, andeven for more remote localities inSiberia.

Efremov rose through the ranks in theSoviet Palaeontological Institute. In1937 he superintended its removal toMoscow when Stalin shifted his seat ofpower from the old imperial capital ofSt Petersburg, where Murchison had

been so lavishly entertained by TsarNicholas nearly one hundred yearsbefore. As well as his Permo-Triassicresearches, Efremov used his position toinitiate the first of a long series of Sovietand Russian expeditions to explore thedinosaur-rich lands of Mongolia.

When Efremov died in 1972, thePalaeontological Institute was stilllocated in the Neskuchny Palace, anarchitectural monument of the eighteenthcentury in the centre of Moscow.Efremov had planned for a larger andmore modern museum, and in 1987, 15years after his death, thepalaeontologists took over a newbuilding on the outskirts of Moscow.The new museum boasted extensive

public displays and a staff of over 100 –an impressive legacy of Efremov’sinfluence and authority. I was able tovisit it in the 1990s, at the beginning ofmy Russian adventures (Chapter 10).

The first complexecosystems

At the same time as Efremov and hisstudents were sorting out the stratigraphyof the Russian Permo-Triassic, SouthAfrican geologists were working on thatof the Karoo, and by 1995 a detailedscheme had been established7 thatallows us to dissect the end-Permian

crisis on land in some detail. We willnot explore all that is known about thelife of each of the zones of the BeaufortGroup in the Karoo Basin, but let uslook at the Dicynodon Zone, just beforethe end of the Permian. In retrospect, ofcourse, we know that this scene ofburgeoning life was not to last, but therecan have been no presentiment of theimpending cataclysm.

If we could visit the DicynodonZone we would experience a time oftropical climates and monsoonal rain.Great rivers cut across the flat plains,sweeping masses of sandstone downfrom the Antarctic mountains to thesouth. Occasionally the rivers spill overtheir banks, forming floods of finer

sediment. Here and there, lakes providea home for fishes. Trees and bushesaround the margins of the lakes droptheir leaves into the waters, where theysink and are entombed in the dark mudsat the bottom. Typical plants around themargins of the lakes and rivers aremosses, horsetails and ferns. Here andthere bush- and tree-like conifersprovide some higher foliage. Aparticularly common plant isGlossopteris, a ‘seed fern’. Usually atree with a woody trunk and standingabout 4 metres tall, some species arelower and more bushy. Glossopterisleaves are tongue-shaped smoothstructures with a central stem, arrangedin star-like bunches of 20 or so at the

ends of branches.Scurrying in the undergrowth are

centipedes, millipedes, spiders,silverfish, cockroaches and beetles.Bright mayflies and dragonflies skip andglint in the dappled sun over the lakes,while bugs and flies feed on the leavesand on animal dung. There are notermites, ants, bees, wasps or butterflies– all of these come much later. Snailsand worms crawl in the mud at the edgeof the lakes, and leave their trailspressed in the wet sediment.

During the dry season, the lakes dryup and water-living animals and plantsdie off. The bugs and beetles can moveelsewhere. The dragonflies and mayfliesdie, but their line is maintained by a few

eggs and larvae in the much-reducedponds. Fishes die too, except for thosethat happen to be in rivers whose watersupply comes from the high, coolmountains to the south. Other riversdwindle to a trickle and disappear,leaving behind an arid, cavernous wadi.

Some animals had strategies tosurvive the hardships of the dry season,burrowing deep in the cool mud andsealing themselves in. It is likely thatlungfishes of Dicynodon Zone timeswere able to do this. They certainly do ittoday, and fossils have been found oflungfishes apparently sealed intoburrows. Sensing that its pool or pond isdrying fast, the lungfish fattens itselfwith all the food it can and then

constructs a chamber as much as a metrebeneath the muddy bed. It then makes awatertight cocoon from mucus and turnsits body systems down to stand-by. Thisprocess of aestivation (‘over-summering’) is analogous to hibernation(‘overwintering’). Some of the Karooreptiles seem to have done the samething: burrows have been found on riverbanks containing clutches of skeletonscurled up and evidently sheltering fromthe heat of the sun. Such fossilizedexamples show us the victims: they hadstarved to death, or perhaps their burrowwas flooded by monsoonal rain beforethey could escape.

It is the vertebrates for which theDicynodon Zone is famous, however.

Seventy-four species are currentlyrecognized, consisting of two species offish-eating amphibians in the rivers and72 species of reptiles. At the base of theplant-eating food chain (Fig. 34) are twospecies of procolophonids. Thesereptiles have triangular-shaped skullsand rows of blunt peg-like teeth thatwere used to crush and tear their plantfood. The procolophonids are probablydistant cousins of modern turtles. At thebase of the flesh-eating food chain arethe millerettids, of which two speciesare known from the Dicynodon Zone.The millerettids look superficially likelittle lizards, but they are probablyrelatives of the procolophonids. Theirtiny pointed teeth suggest a diet of

insects. Other insect-eaters include aspecies each of Youngina andSaurosternon, both of which alsolooked like lizards but were actuallymuch more primitive. There are alsothree species of small, possibly insect-eating therapsids.

34 A simplified food web of themain reptiles of Dicynodon Zonetimes, one of the first complexecosystems on land and entirelydestroyed by the end-Permianmass extinction event.

Next up in size are thetherocephalians, of which 15 specieshave been reported so far from theDicynodon Zone. The therocephaliansare cat-sized and have longish skullswith sharp, piercing teeth concentratednear the front. In the same dietarycategory are the early cynodontsProcynosuchus and Cynosaurus, whichare the size of terriers and have a full setof teeth along the length of their jaws.These were all probably second-level

predators that hunted the smaller plant-eating procolophonids as well as theinsect-eating anapsids, diapsids andsynapsids.

The Dicynodon Zone is named, ofcourse, for Dicynodon, of which ninespecies are recognized; a further ninespecies of other dicynodonts are alsoknown. These toothless herbivores, ofthe type first described by Owen fromSouth Africa in 1845, had clearlyundergone a great diversification. Thedifferent species varied somewhat insize, as Andrew Geddes Bain hadnoticed when he made his firstcollections. Doubtless each species fedon a slightly different plant diet,allowing such a diverse array of 18

closely related species to co-exist, justas numerous species of apparentlysimilar antelope do on the grassysavannahs of Africa today.

The largest herbivores include twospecies of pareiasaurs. Hefty animals,each about 2 metres long and withmassive, barrel-shaped bodies carriedon rather feeble-looking sprawlinglimbs, these reptiles are close relativesof the much smaller procolophonids.Pareiasaurs are renowned for theiruncomely looks, having broad mouths, aroughly triangular skull and numerousprominences and excrescences all overthe head. Their unfortunate plainness ofvisage was matched by low intelligence,as indicated by the relatively small size

of the head in comparison to the body,and the even smaller size of thebraincase within the skull. Theydoubtless munched placidly onGlossopteris leaves, and resisted attackwith their formidable size and knobblyhead.

The dicynodonts and pareiasaurs arepreyed on by a fiendish array ofgorgonopsians, some 25 species of them.These predators (Fig. 35), sometimessmaller than their prey, being generallyabout 1 metre in length, were by nomeans ineffective. They all haveexcessively elongated canine teeth, aswell as the jaw mechanism and musclesto operate as sabre tooths, a mode ofattack that no longer exists. Until the end

of the last ice age, 10,000 years ago,various sabre-toothed cats, with curvedknife-like canine teeth up to 15centimetres long, hunted the large thick-skinned mammoths, mastodons andwoolly rhinos of Europe and NorthAmerica. They leapt on to the backs oftheir much larger prey and sank the teeththrough the skin. By twisting or tearing,they caused a fatal wound. Thegorgonopsians of the latest Permianprobably employed a similar strategy:piercing, tearing and waiting for theirdicynodont or pareiasaur prey to bleedto death.

35 The gorgonopsian Inostranceviafrom the latest Permian of Russia.

The key fact about the DicynodonZone is the ecological maturity of thevertebrate fauna. Food webs (see Fig.34) were complex, with herbivores ofvarious sizes feeding on a broad rangeof plants, and with three or more levelsof predators feeding respectively on thesmall, medium and large herbivores.This kind of complexity and diversity ofland vertebrate life is common enough

today, but was first achieved only duringthe Late Permian. Before that time,herbivorous reptiles were rare or non-existent, and flesh-eaters existed only ina couple of size categories. The tertiary-level sabre-toothed gorgonopsians weresomething new altogether.

Collapse of stout pareiasaur

All this complexity was wiped out bythe end-Permian crisis. At the moment,we can only see the picture partially, asif through several layers of frosted glass.But the astonishing finding from all thecurrent work is that perhaps the end-Permian crisis was actually more severe

for life on land than for life in the sea.This is a striking turn-around from theview of only 15 years ago, thatessentially nothing out of the ordinaryhad happened on land at all.

It is hard to work out the precisepatterns of collapse of the plants andinsects because of their poor fossilrecords in the Karoo. However,worldwide, 11 families of true landplants fell to three across the Permo-Triassic boundary. Species numbers fellfrom about 140 to 50 or fewer: thequality of the evidence is just not goodenough to be sure. The story appears tobe clearer for the vertebrates.

The 74 species of amphibians andreptiles from the Dicynodon Zone in the

Karoo contrast with only 28 in thesubsequent Lystrosaurus Zone, theearliest phase of the Triassic. But onlytwo species of reptiles have beendiscovered that survived through thePermo-Triassic boundary, thed i c yno d o nt Lystrosaurus (speciesundetermined) and the therocephalianMoschorhinus kitchingi. The other 26species of the Lystrosaurus Zone had allarisen during the intervening period. So,at species level, the end-Permian crisisapparently accounted for 72 of the 74vertebrates, a loss of 97%. But we knowthis is an overestimate.

A number of reptile groups areknown to have survived into theTriassic, such as the procolophonids, the

basal diapsids, the dicynodonts and thecynodonts, so at least a further threespecies must have passed through thePermo-Triassic boundary. Perhaps theirfossils have yet to be found, or they mayindeed have died out in the Karoo Basinbut survived elsewhere, and thenrecolonized South Africa in the earliestTriassic.

Worldwide, the picture seemsclearer.8 Of the 48 families ofamphibians and reptiles that werepresent in the last 5 million years of thePermian, 36 died out, representing a lossof 75%. This is higher than the generallyaccepted figure of a 50% loss offamilies in the sea. So, for thevertebrates at least, the end-Permian

crisis was apparently more severe thanfor marine animals. If a 50% family lossscales to 90% species loss in the sea,then the loss of 75% of vertebratefamilies must scale to something like95% species extinction on land. Whetherin South Africa, or worldwide, thepareiasaurs, millerettids andgorgonopsids, as well as numerous othergroups, had gone for good.

New work in the Karoo Basin hashelped to identify just what theenvironmental changes were on land.Roger Smith, a sedimentologist from theSouth African Museum in Cape Town,and Peter Ward, a palaeontologist fromthe University of Washington, Seattle,pinpointed the exact position of the

Permo-Triassic boundary in the BeaufortGroup sequence.9 The boundary betweenthe sediments of the Dicynodon andLystrosaurus zones was marked by a20-metre thick succession of channelsandstones and finely bedded thinsandstones and shales. The end of theDicynodon Zone is marked by the lastappearance of fossils of Dicynodon andof Glossopteris, and this also marks thebeginning of the Lystrosaurus Zone.There is not a single level at which allthese things happen. In fact, the lastDicynodon occurs at the top of thesection, while the first Lystrosaurusoccurs at the base, so there is anoverlap. However, the two dicynodontsare generally found in different beds.

The last specimens of Dicynodonare found preferentially in green- andolive-coloured mudstones and finesandstones, while Lystrosaurus is foundin red-coloured sediments. In the largerpicture, the green-coloured sediments ofparts of the Dicynodon Zone arereplaced by red beds. So, Smith andWard interpret this to mean that therewas a step-by-step change in theconditions of deposition of thesediments, which is seen in detail in the15-metre transitional zone. Just whatwas going on?

Plant die-back and

catastrophic erosion

The red and green sediments ofDicynodon Zone times were depositedby broad meandering rivers, carryingsediment at relatively steady rates fromthe southern mountain ranges. Sedimentsin the lower parts of the LystrosaurusZone are quite different, consisting ofmore sandstones and less mudstones.The sandstones are not seen in discretechannel fills but form instead thicker,more complex packets of sediment,evidently deposited by braided rivers.Braided rivers, where the individualchannels twist and turn dramatically, aretypical of alluvial fans, formed whenfast-flowing, high-energy upland parts of

river systems come hurtling down themountains and meet a lower-angle slope.

Smith and Ward argue that a majorchange in sedimentation patterns washappening in the Karoo Basin just at thePermo-Triassic boundary. They believethat the pace of sedimentation increaseddramatically. In the 15-metre transitionalsection, the Dicynodon-bearing greenbeds of the lowlands are progressivelyoverstepped by higher-energy redsandstones with associatedLystrosaurus, until the latter associationof sediments and fossils prevails.

What caused the increase in the rateof sedimentation? One obvious causemight be that the mountains that suppliedthe sediments had been uplifted by some

major tectonic activity. However, thereis no evidence for anything of the kind.So, Smith and Ward propose that theincrease was caused by a major die-back of plants. Plants prevent erosion onland by binding rocks with their roots,and building up layers of soil. If allplants are removed from any typicalcontinental area, rates of erosionmultiply by 10 or 20 times – as seen inthe catastrophic effects of deforestationon the foothills of the Himalaya in recentyears. After the trees had been removedfor building huts and for firewood,rainfall washed huge amounts of soil androck from the slopes, causingcatastrophic floods in Bangladesh everyyear or so.

Smith and Ward only have goodevidence for this major change insedimentation in the Karoo Basin, butthey believe that other continentalPermo-Triassic boundaries show asimilar phenomenon. So perhaps thiswas a worldwide aspect of the end-Permian crisis. Their hypothesisdepends on the dramatic loss of plantlife over wide areas. What evidence isthere?

Detailed studies of ancient soils, andof plants, across the Permo-Triassicboundary seem to support the Smith-Ward hypothesis. Greg Retallack of theUniversity of Oregon has studiednumerous Permo-Triassic soilsuccessions in Gondwana (Antarctica,

Australia, New Zealand), and he finds ashift in soil and plant types.10 LatestPermian soils include coals and rootedsands, while earliest Triassic soils aremainly root-filled mudstones. Thisindicates a shift from cold-temperatebroad-leaved swampy floras, dominatedb y Glossopteris, ferns, horsetails,mosses and the like, to cool-temperateconifer forests. Temperatures and levelsof rainfall rose slightly across thePermo-Triassic boundary, and this wasassociated with increased erosion andrunoff.

The disappearance of swampvegetation for a time at the beginning ofthe Triassic is associated with thefamous ‘coal gap’. Nowhere in the

world did coals form during the 20million years of the the Early and MidTriassic, whereas coal is usuallyforming somewhere, wherever climatesare warm and humid. Early Triassicfloras were dominated by stress-tolerant, opportunistic kinds of plants –what are commonly called ‘weeds’ –which are adapted to spread quickly andto live in poor-quality soils lacking innutrients. These are the plants that willgrow up first on a piece of disturbedground, such as the rose-bay willowherband other weeds that colonizeabandoned building sites. In time theweeds give way to different plants,representing the more maturesuccessional flora.

Some additional evidence tends tosupport this idea of a catastrophic lossof plant cover in the earliest Triassic.When he was examining samples ofsediment under the microscope, YoramEshet, a palaeontologist from theGeological Survey of Israel, noticed atremendous shift in pollen and sporetypes across the Permo-Triassicboundary.11 In the latest Permian, insamples taken from all over Israel, hefound a typical selection of pollen andspores from a wide range of terrestrialplants. The same was true for samplestaken from higher levels in the EarlyTriassic. But for a few metres at the verybeginning of the Triassic, almost alltypical plants disappeared, and all he

could see were cells of fungi. In otherwords, the end-Permian crisis wasmarked by a fungal spike, a dramaticspread of mushrooms all over Israel.

Not only Israel, but the world. Eshetfound that other authors had spotted thefungal spike in Europe, North America,Asia, Africa and Australia. It has notbeen observed in Greenland, probablybecause the sections there are so thickthat an increase in proportions of fungi isspread through several metres. So,something had temporarily displaced thenormal floras and encouraged fungi toblossom, or at least to extend theirhyphae and spores widely. Eshetsuggests that the sudden proliferation offungi indicates a high-stress event that

killed most other plant life. Fungiflourish in the presence of decayingorganic matter, and the widespreadkilling of more complex plants provideda perfect environment for them. But theplants had not been entirely wiped out.Some clearly survived in isolatedpatches. When the stress was removed,the normal diversity of plants returned,and the fungi retreated to their usualposition in the system.

Plant ecosystem collapse

In a recent study, Cindy Looy of theUniversity of Utrecht and RichardTwitchett, then of the University of

Southern California,12 were able to teaseapart in some detail the step-by-stepprocess of the plant ecosystem collapse.They showed quite convincingly thatplants did not disappear overnight, andthat normal ecological processescontinued for some time, even during thephase of severe environmental stress.They based their work on the remarkablePermo-Triassic succession inGreenland, which we encountered inChapter 7 – the transition from the latestPermian Schuchert Dal Formation to theearliest Triassic Wordie CreekFormation. These sediments weredeposited in shallow seas and are wellknown for their marine fossils. But theywere also close enough to shore to

contain abundant pollen and spores thathad been blown offshore from landplants. So there is a unique opportunityhere to see what was happening in thesea and on land, and to match the timing,centimetre-by-centimetre, through thesediments.

This is the story of the plantextinctions. The first phase saw the lossof closed woodland to open herbaceousvegetation – more bushes than trees.Fungal activity increased at the sametime, possibly related to the decay of thetrees that were dying off. The place ofthe trees is taken by opportunistic,weedy species – smaller and faster-growing – which take advantage of thegaps in the once uninterrupted forests.

These species of weedy clubmosseswere previously kept in check by thedominance of woody trees, the seedferns and conifers, which blockedsunlight from the forest floor. With thetrees gone, the small green plants thatwere formerly suppressed nowexpanded and spread.

In a second phase of the collapse, theopen space was further occupied byferns; these had been present before, butin low abundance. In addition, some newforms arrived from the south. Cycadsand other plants that had been knownbefore only in North America, forexample, suddenly appear in the crisisfloras of Greenland. So, in the face ofdeath and destruction, it seems

opportunistic species colonized and tookadvantage of the crisis. But not for long.

Then, in the final killing phase, mostof the invaders disappear. The last of thewoody trees and bushes go too. All thatis left is a much-reduced flora oflycopsids, club-mosses – low plants thatpresumably survived in damp pocketshere and there. During the three-phasekilling cycle, the overall volume ofpollen and spores diminishesdramatically, indicating rarity of plantlife at the final killing level, and formany metres of sediment thereafter in theearliest Triassic.

How long did all this take? Looy andTwitchett estimate that the three-phasedecimation of floras lasted for some

500,000 to 600,000 years, spanning thePermo-Triassic boundary,corresponding to the duration of marinedecline in the Meishan section in China.

Lystrosaurus the survivor

What of the reptilian survivors of thecrisis on land? The most prominent wasLystrosaurus, the modest-sizeddicynodont which we met in Chapter 1and which, as we have seen, overlappedfor a short time with its typical latestPermian relative Dicynodon. Theextraordinary thing about Lystrosaurusis not that it survived, but that itdominated the whole world for a short

time. Species of Lystrosaurus have beendescribed not only from South Africa,but also from South America, Antarctica,India, China, Russia and possiblyAustralia. The Lystrosaurus Zone ofSouth Africa is matched bycontemporary rock units in all thosecontinents. The apparent absence ofLystrosaurus in other areas is probablydue to the fact that the right rocks haveyet to be found. It was trulycosmopolitan – a time indeed when‘pigs ruled the earth’.

Lystrosaurus was not onlywidespread but it was also hugelyabundant. In South Africa,palaeontologists cry with frustrationwhen they find another Lystrosaurus

skull. James Kitching, one of the greatestcollectors in South Africa, estimates thatover 2000 skulls of five species ofLystrosaurus have been collected in theKaroo Basin, and many more have beensmashed with hammers. Collectors aredesperate to find anything else. But thereptiles and amphibians associated withLystrosaurus are unbelievably rare.Only a few specimens of its smallerrelative, the dicynodont Myosaurus,have been found. The same is true for theo t h e r Lystrosaurus Zone animals:isolated specimens of a few species ofamphibians, a couple of procolophonidsand basal diapsids, and 10 species ofsmall flesh-eating therocephalians andcynodonts. So, in all, Lystrosaurus

makes up at least 95% of all animalsfound in the Lystrosaurus Zone.

Such dominance by a single animalis far from normal. If the 20 or so otherspecies comprise less than 5% of thefauna, we are looking at something thatis so unnatural as to resemble a productof human intervention. For example, afield of cows is a typical agriculturalmonoculture. True, there may be half-a-dozen rabbits in one corner of the field,and a few field mice, voles and weaselslurking in the hedgerows, but the sceneis dominated by cows. Suchmonocultures cannot survive withouthuman assistance. In Nature, when foodis abundant, the one sure thing about lifeis its diversity. No one species will

dominate, and there may be as many asfive or ten that are considered to beabundant.

The unnatural abundance ofLystrosaurus confirms that it is adisaster species, something like a weedyplant. However, unlike the fungal spikeand the weedy plant species in theearliest Triassic, ‘normal’ faunas couldnot re-establish themselves. Among theplants, sufficient diversity had evidentlysurvived here and there to allow normalfloras to re-establish themselvesrelatively rapidly in the Early Triassic.But Lystrosaurus was on its own. Atabout 1 metre in length, it was the largestanimal on Earth. All the others weremuch smaller. This was a huge change

from the latest Permian situation, whenmany species of dicynodonts,pareiasaurs and gorgonopsians werelarger. But they had all died out.

What did Lystrosaurus do? Like alldicynodonts, it had two upper canines,with otherwise toothless jaws no doubtlined with a horny beak. It used thestandard dicynodont feeding mechanismof grasping a mouthful of leaves at thefront of the beak, perhaps using thecanines as rakes to gather the plantmaterial into a bundle, then closing itslower jaws and pulling them back,slicing the plants into bits. The roughlycut stems and leaves were thenswallowed more or less whole, sinceLystrosaurus could not chew by side-to-

side movements. Doubtless much of thedigestion of the tough plant material tookplace inside the huge gut. Like mostherbivores, Lystrosaurus had a vast ribcage and enormous guts (plant food takesa lot more digesting than meat). Thegurgles and eructations would have beenan Early Triassic phenomenon.

When Lystrosaurus was named in1903 by Robert Broom, he classed it asan aquatic animal, an active swimmer,perhaps something like a pygmyhippopotamus.13 This view held swayuntil the 1990s, when Gillian King andMichael Cluver of the South AfricanMuseum in Cape Town pointed out thatLystrosaurus was actually rather likeother dicynodonts, and was almost

certainly a land-dweller. In oppositionto this, Russian palaeontologist MikhailSurkov of Saratov University hasrecently argued that the Russian species,Lystrosaurus georgi , at least, andpossibly other species of Lystrosaurus,did have adaptations for swimming. Theforelimbs are relatively larger than inother dicynodonts, and the distribution ofthe muscles indicates that the animalused these to paddle vigorously, whilethe hindlimbs trailed behind. Of course,it was still as adept on land as any otherdicynodont, which isn’t saying much.

A superior pig-like reptile?

So why did Lystrosaurus do so well? Itis easy to ask such a question, sincethere is no doubt that Lystrosaurus wasin its way one of the most successfulreptiles of all time, known from a dozenor more species from all continents, andsurviving for a few million years at thebeginning of the Triassic. But perhaps itis wrong to seek particular reasons whyLystrosaurus was so successful. It wasa success faute de mieux. It may havebeen no more than pure chance that thisgenus, among all the others of the latestPermian, did not die out. Then, as thesole modest-sized survivor, it had anempty world all to itself. PerhapsLystrosaurus was actually notparticularly skilled at surviving: it just

had to be good enough to scratch aliving. There were no other significantplant-eaters with which to compete, andthere were no large predators to pursueit.

This sort of question is very popular,as we saw in Chapter 1. But survivingan extinction is not like winning asprinting race. In the case of the race, theathletes know in advance exactly what isexpected. They train and try to hone theirperformance to the very best they can.And, on the day, the best athlete wins.An extinction event is like a race withunpredictable obstacles and of unknownlength. The athletes have no advancewarning of the obstacles, any of whichcould be fatal. So in this race, the

athletes all have pretty much the samechance of reaching the other end of thetrack first – if at all. This is alsoevidently true of the mass extinctions inthe past.

Back in the USSR

The Karoo Basin has yielded up some ofits secrets. Thanks to the work ofBroom, Kitching, Smith, Ward and manyothers, we now know something aboutthe nature of the end-Permian crisis onland. The rate of extinction amongreptiles was as high as any extinction inthe sea, and older ideas that not verymuch happened at the Permo-Triassic

boundary among continental organismshave been thoroughly rejected.

The Russian Permo-Triassicsequences are just as good, and just asrich in some places. In the early 1990s,before Roger Smith and Peter Ward hadpublished their work on the Karoosuccessions, there was an opportunityfor me to go to Russia. The communistregime had collapsed, and Boris Yeltsinhad come to power. Russian and westernscientists were looking for contacts, andI met Akademik Leonid P. Tatarinov, anoted expert on the Russian mammal-like reptiles, but now retired, at aconference in Montbéliard in France in1992. Tatarinov suggested that theUniversity of Bristol should collaborate

with the Palaeontological Institute inMoscow, and we signed an agreement.

This was a very exciting opportunity.We knew the older published work byEfremov and others. Surely Russia mightprovide a good testing ground forhypotheses based on the rock sequencesin the Karoo Basin? Surely also it wouldbe useful to look at a part of the worldthat in the Permian was a long way fromSouth Africa? Are the extinction patternscomparable, or are there some regionaldifferences? Many of the Russianreptiles are very similar to those fromSouth Africa, but there are some uniqueRussian groups, not known fromelsewhere. So could some of theextinction patterns be different? What of

the apparent loss of plant cover andheightened erosion that seemed to occurin the Karoo: could this be confirmed inRussia as well?

It was with these questions in mindthat we went to Russia in 1993.

10

ON THE RIVER SAKMARA

The dragonflies flew above our heads atdusk, swooping and weaving in pursuitof mosquitoes. They did not appearduring the day, but emerged at nightfall,when their prey came out. They wereshadowy forms, with their great pairedwings, more like birds than insects.

‘Ubivayoot kameri. Harashoa!’murmured one of our Russian

colleagues, ‘They kill the mosquitoes.Good!’.

The bonfire crackled in front of us,and we crowded round in an attempt toescape the biting insects. This was theroutine every night. We ate duringdaylight, and spent the early eveningtrying to avoid being eaten by thevoracious blood-sucking mosquitoes.They were larger than anything we wereused to at our respective homes inBritain and North America, and all ourlotions proved totally ineffective againstthem. The leader of the trip, ValentinTverdokhlebov of the GeologicalInstitute in Saratov, offered us someRussian anti-mosquito mixture. This wasa pinkish lotion, rather like baby cream,

with an unusual smell. We never foundout what was in it, but it certainlyworked.

We were here, by the SakmaraRiver, at the south of the Ural Mountainsin Russia, in July 1995. The Sakmaraflows southwestwards from the foothillsof the Urals to join the Ural River itself,which then passes southwards to theCaspian Sea. Our camp site was not farfrom some of the old Copper Sandstonemines that had produced fossils in theeighteenth and nineteenth centuries.These had long been out of action, andwe hoped to find specimens in the redsandstones and mudstones of thesurrounding foothills and river banks.

Our aim was to explore the Russian

Permo-Triassic succession in detail tosee whether it might shed light on whathad happened 251 million years ago.

In the galleries

Two years earlier, during our first visitto Russia, Dr Glenn Storrs and I spent aweek at the Palaeontological ResearchInstitute (known as PIN,Paleontologicheskii Instituta Nauk) inthe outskirts of Moscow, on the avenueProfsoyuznaya. This was the instituteIvan Efremov had dreamed of, but neversaw in his lifetime. Glenn was workingthen at the University of Bristol, and heis now Curator of Fossil Vertebrates at

the Cincinnati Museum of NaturalHistory.

We stayed at a hotel belonging to theRussian Academy of Sciences in centralMoscow. When we arrived there wewere given 10,000 roubles each,seemingly a huge sum, but worth onlyabout £70 then. Andrey Sennikov, one ofour Russian colleagues, explained thepurpose of the money: ‘This is forscientific purposes. You must not spendit on bad girls.’ We heeded Andrey’sadvice, and spent the week studying theremarkable collections of fossil reptilesat PIN.

PIN is both research institute andmuseum. In its heyday, in the last yearsof the Soviet Union, and for a few years

after perestroika, and Gorbachev andYeltsin, PIN employed around 100palaeontologists, both in an olderbuilding in central Moscow and in thenew museum that had opened in 1987 onProfsoyuznaya. Numbers have dwindledsince then, as salaries have gone unpaidand people have been forced to leave foreconomic and other reasons. The newmuseum is a striking brick edifice, withan extensive basement in which vastnumbers of bones are stored, and twofloors of public galleries. The publicgalleries are a wonder. The architect hasused the structure and ornament of thebuilding to illustrate its subject matter.There are great terracotta sculptures ofmammal-like reptiles, dinosaurs and

mammoths on the walls, ceramic imagesof pterosaurs, coelacanths and ancienthumans, and a trompe-l’œil stratigraphiccolumn, using mirrors, which gives astriking impression of the seemingendlessness of geological time as youlook down. The galleries at PIN are fullof wonderful displays of theantediluvian treasures of the SovietUnion, including some of the earliestforms of life from the Precambrian ofSiberia, as well as dinosaurs and fossilmammals from Mongolia. Glenn and Iwere there to look at the Permo-Triassicreptiles and amphibians.

In the very first Permo-Triassicgallery we were astounded. There, as acentrepiece, was an assemblage of 20

pareiasaur skulls and skeletons.Pareiasaurs (Fig. 36) had beendescribed first from the Karoo Basin, aswas noted in the last chapter, and theywere later found in abundance in Russia.From quarries near Kirov, endlessnumbers of specimens of the pareiasaurScutosaurus have been excavated andbrought back to Moscow. Russianscholars had divided their pareiasaursinto some 15 species, but later work hasshown that this was more to do withnationalistic pride than reality. Therewere probably in reality only twospecies.

Ranged around the walls of thePermo-Triassic gallery were skeletonsof many other kinds of amphibians and

reptiles, all from the southern UralMountains, from northern parts of theUrals, and from Arctic Russia, north ofMoscow, especially along the NorthDvina River – the fruits of Amalitskii’sexcavations of the early twentiethcentury. Most striking were thesynapsids, which are very similar tothose from the Karoo. These range fromsmall insect-eating cynodonts andtherocephalians, to large pig- and hippo-sized dicynodonts. The most astonishingof the Late Permian synapsids was thegorgonopsian Inostrancevia. Its largewolf- or tiger-like skeleton indicates apowerful runner; its long dog-like skullsuggests intelligence. The canine teethare arresting – fully 50 millimetres long

– and the lower jaw could drop open toan angle greater than 90°. Thisgorgonopsian was similar to thoseknown from the Karoo.

36 The Russian pareiasaurScutosaurus from the latestPermian: top, skeleton; Below left,skull viewed from above and,below, right from the front.

37 The latest Permian Vyatskianfauna from Russia. At the back, thegorgonopsian Inostrancevia looksspeculatively at the plant-eatingpareiasaur Scutosaurus. Adicynodont stands at the water’sedge, while the flesh-eatingsynapsid Annatherapsidus sits on alog, with Dvinia below. Thetemnospondyl amphibianChroniosuchus sits on a sand bank,with Kotlassia in the water. In theforeground, the littleprocolophonid Microphon is to theleft, the temnospondyl Raphanodonto the right.

Here, in the Late Permian of Russia,is an assemblage very like that of theDicynodon Zone of South Africa, which

we examined in the last chapter. Thelatest Permian fauna of Russia, theVyatskian assemblage (Fig. 37), knownfrom the North Dvina river and from theSouth Urals, was rich and diverse: thepareiasaur Scutosaurus, the dicynodontDicynodon, four species ofgorgonopsians, including Inostrancevia,and two smaller carnivores, atherocephalian and a cynodont. In otherlocalities, latest Permian reptilesinclude Archosaurus, a 1-metre longslender fish-eating reptile, oldestmember of the Archosauria, or ‘rulingreptiles’, the group that includescrocodiles and dinosaurs, andprocolophonids, small triangular-skulledreptiles, related to pareiasaurs, but

superficially looking somewhat like fatlizards. At the water’s edge were threeor four species of amphibians.

This was a rich and complexecosystem, with as many animals as inany modern terrestrial community. Therewere herbivores specializing in plants ofdifferent kinds, fish-eating amphibians,insect-eating synapsids, carnivoresfeeding on small prey, and thegorgonopsians, so-called top predators,feeding on the largest of the herbivores.All these animals were wiped out by theend-Permian crisis. But what happened?We had to learn more about the rocksfrom which the bones came.

On the banks of the Volga

During our first visit to Moscow, wewere able to visit the famous Tikhvinsksite, near the town of Rybinsk. Wereached Rybinsk by train. Rail travel isthe main means of travelling longdistances in Russia, and the trains areclean and efficient, if not particularlyfast. We packed into a sleepingcompartment for four: Glenn Storrs,myself, Andrey Sennikov and IgorNovikov, who was Deputy Director ofPIN. The trip was organized as part of ajoint research scheme, funded by theRoyal Society (London) and the RussianAcademy of Sciences (Moscow).

The rail tickets were cheap, and we

had invested in a second-class berth.Third-class passengers slept in opencarriages divided by bunks into sections,sleeping people on three levels. Ourcompartment was very like a British orFrench sleeper, with a two-level bunk.We settled on the train in Moscow, andprepared to eat. Andrey brought outsome small, round, mid-white loavesand rye bread, dried river fish and freshcucumbers. The train attendant brought achina teapot and mugs, which we hiredfor a few roubles. Hot water came froma fiendish small boiler set in an alcoveat one end of the carriage, which wasalways kept hot. We feasted on sweettea, bread and fish. We then sleptthrough the slow rocking journey, and

awoke at six in the morning to findourselves in Rybinsk.

Rybinsk lies at the confluence of theVolga and Sheksna rivers, just belowRybinsk Reservoir, a huge artificiallake, over 120 kilometres long, createdin the 1940s by damming the two riversto produce a hydroelectric powerscheme. We travelled a few milesdownstream on the Rackyet (‘rocket’),an astonishing Russian-built hydrofoilthat plies up and down the small townsof the Volga, touching the shore on eitherside at landing stages, until we reachedour destination, Tikhvinsk.

The Tikhvinsk site was famous asthe source of some beautiful skulls ofamphibians, the first of which had been

collected by Ivan Efremov in the 1930s.The amphibians of the Permo-Triassicwere mainly temnospondyls, a group thatprobably contains the ancestors ofmodern frogs and salamanders. As inmodern frogs, the temnospondyls hadbroad low skulls, with a wide curvingfront to the snout and jaws lined withdozens of modest sharp teeth, ideal forsnapping at fishes. The temnospondylsprobably lay in wait on the beds ofstreams and lakes, or in the shallowsamong plants, and engulfed their prey bysuction. Opening a broad mouth rapidlyunder water creates a region of lowpressure inside, and any fish, or otherdebris, nearby is sucked in. Thebraincase was tiny and the cranium and

the lower jaw were of similar depth,again as in modern frogs andsalamanders. They could open their jawby raising the skull, an ideal arrangementfor suction feeding on the bottom.

38 The earliest TriassicRybinskian fauna from Russia, asseen at the Tikhvinsk locality. Inthe background the basalarchosaur Chasmatosuchuswanders past the dicynodontLystrosaurus. In the centre aresome small reptiles: atherocephalian (left) looks at theprocolophonid Tikhvinskia and theprolacertiform Boreapricea, whichhas an insect in its mouth.Amphibians include the heftytemnospondyl Wetlugasaurus on thebank, and Benthosuchus swimming.

The sediments at the Tikhvinsk siteare classified by Russian stratigraphersas part of the Rybinskian Gorizont, orhorizon, a unit of rocks laid down in the

Early Triassic, perhaps three or fourmillion years after the great extinction.The Rybinskian Gorizont can be tracedover the Moscow region, and equivalentunits occur in the southern Volga andUrals regions. The shales contain plantremains and molluscs that point todeposition in rivers and lakes, butperhaps with some marine influencefrom an Early Triassic sea in the Balticregion.

The striking thing about theRybinskian fauna (Fig. 38) is howdepleted it is in comparison with theLate Permian Vyatskian assemblagerepresented by the great collections ofpareiasaurs, dicynodonts, cynodonts andother reptiles we’d seen in the Moscow

museum. The commonest animals are thetemnospondyl amphibians Benthosuchusa nd Thoosuchus, each 0.7–1.5 metreslong. Much rarer is the archosaurChasmatosuchus, which might havepreyed on the temnospondyls, or onfishes, and the small, superficiallylizard-like procolophonid Tikhvinskia.The dominant Permian group, themammal-like reptiles, are represented inthe Rybinskian fauna by just one smallform, Scalopognathus, which is actuallyknown from only a single jaw. Otherreptiles have been reported by Russianauthors, but the remains are partial andequivocal.

This is exactly what we saw in theKaroo Basin. The complex latest

Permian ecosystems of the DicynodonZone disappeared suddenly, and theearliest Triassic Lystrosaurus Zonefaunas were monocultures of thedicynodont Lystrosaurus, with a few,extremely rare, small reptiles inassociation. The Russian situation wasrather different – no Lystrosaurus atTikhvinsk, and mainly amphibians. Butperhaps this was caused by theenvironments there, the deposits of blackshales suggesting possibly a deep lake.Or perhaps Lystrosaurus was alreadyextinct. The Russian Lystrosaurusgeorgi comes from the VokhmianGorizont, at the very base of theTriassic, and a few million years olderthan the Rybinskian Gorizont at

Tikhvinsk.

First sight of the Sakmara

Our appetites whetted by this first visit,Glenn Storrs and I were keen to see theclassic bone-beds of the Urals. ThePermo-Triassic of the Moscow basinwas good, and it had produced excellentfossils over the previous centuries, butlocalities are isolated and small,surrounded by vegetation, farms andtowns. Geologists prefer deserts, coastsand ravines, where the rocks areexposed. We arrived in July 1994 on oursecond visit before heading for theUrals. We went at the invitation of

Valentin Tverdokhlebov and VitalyOchev in Saratov. Ochev is adistinguished palaeontologist who hasworked on the fossil amphibians andreptiles of the Permo-Triassic of theUrals since the late 1950s. Glenn and Iknew him by reputation, since he hadpublished widely in both Russian andinternational scientific journals.Tverdokhlebov was less known to us.He is a field geologist, specializing insedimentology – the study of ancientsediments and the interpretation of theevidence they provide aboutenvironments and climates. He knowsevery inch of the South Urals area sincehe has worked for the Geological Surveyof the USSR, and now of Russia, for

decades, producing geological maps andsearching for valuable minerals andores.

Andrey Sennikov took us to the greatKazan Station in Moscow where wewere to board our train to Orenburg. Thetrain left about midnight, as do mostlong-distance trains in Russia. We hiredour teapot and bedding, as usual, andturned in for the night. By morning, wehad passed through Lyubertsy, Kolomnaland Ryazan, and were well out on thegreat plain to the southeast of Moscow.Birch forests and farmland extended forhundreds of kilometres in all directions.By early afternoon, we had passedSyzran, and had our first sight of theVolga. The railway track crosses the

Volga at a narrow point, but even so thebridge is nearly 3 kilometres long. Theriver is so wide that the other bank is notvisible. We soon arrived at Samara, onthe east bank of the Volga, an industrialcity of over a million inhabitants, and amajor administrative centre. Factoryblocks proclaimed ‘Slava Oktyabr’skoiRevolyutsii’ (‘long live the OctoberRevolution’). This city, like so manyothers, had recently changed its name:older atlases show it as Kuybyshev.

In Samara, as we waited for peopleto get off in the afternoon sun, we sawanother train pulling in from the south,from Tashkent, the capital of Uzbekistan,one of the smaller Central Asiaticrepublics. As the doors opened, large

numbers of huge melons rolled on to theplatform. In fact, the third-classcarriages were entirely full of melons,on the floor, stacked on the beds and inthe corridors. The small brown-skinnedUzbek farmers in their long dark robesand woollen hats, and their families,rolled out on to the platform, followingtheir fruit. We understood from ourcompanions that it was economicallyviable for them to buy cheap tickets tomake the two-day journey into the heartof Russia to sell their melons. And itwas a great deal safer than entrustingtheir produce to commercial carriers.

Eventually, after a 30-hour journey,and a second night on the sleeper, wedrew into Orenburg Station. Valentin

Tverdokhlebov and Vitaly Ochev, andtheir crew, were there to meet us.Tverdokhlebov is a tall, thick-set manwith dark hair and a tanned face, clearlymore used to fieldwork than to theoffice. He spoke only Russian, but Imanaged to make myself understood.Glenn amazed the Russians by speakingto them fluently in their own tongue: hehad learnt Russian at school, largelybecause of his great passion for Russianchoral singing. Ochev hopped abouteagerly, speaking a mixture of Englishand German. Like many older educatedRussians, he had learnt one or morewestern languages as a child. It was hardto remember that Ochev was over 70,but still hugely active. The remainder of

the crew were there, and one of them,Leonid Shminke, spoke good Englishand he was to act as our liaison. Wepiled into a Waz minibus and set off.The Waz minibus is ubiquitous andalways khaki. The name Waz (properlyUAZ) stands for Ul’yanovskiiAvtomobil’nyi Zavod (Ulyanovsk CarFactory). These buses carry nine people,and they look very like a classicVolkswagen camper of the 1960s.

We headed out of Orenburg,travelling eastwards through the villagesof Sakmara, Chorny Otrog to KolhozPravda (‘collective farm truth’). Afterthe village of Gavrilovka we passed onto a well-made graded dirt road, orgraida, and then shot off to the side and

through farm fields to the banks of theSakmara. There, among the trees,Tverdokhlebov and his colleagues hadbuilt a magnificent camp site. Five or sixsubstantial tents were ranged round aclearing. The Sakmara at this point isabout 20 metres wide, flowing stronglybetween high banks. The Russians hadcut neat steps down the side and hadbuilt a small jetty. Then we heard acheerful cry of ‘What on earth are youdoing?’ from a boat, and an old Russiancame ashore with a bucket full of fish.This was Peter Tchudinov, a notedpalaeontologist, expert on synapsidreptiles of the Permo-Triassic. But hewas retired now, and only here for thefishing, as he made clear in his perfect

English.

Getting to grips

We had just over a week on the banks ofthe Sakmara. Evidently, the Saratovcrew came to this spot each year. In theold, Soviet days, Tverdokhlebov hadbeen commissioned to complete a map,or search around the area for somespecific minerals. The whole southernUrals region is rich in iron, copper, zincand uranium, as well as oil and gas, andgeologists have been enormouslyimportant in developing major industriesin the area. We now looked forward tosome field work. We were probably the

first westerners in the twentieth centuryto see the classic Russian fossil reptilelocalities. Some of them had beenvisited, briefly, by Murchison in 1841and by Twelvetrees in the 1880s, butfree travel in remote parts of Russia byforeigners was virtually impossibleduring Soviet times.

Our intention was to try tounderstand the succession of animals inthe Permo-Triassic rocks of the Urals.We had read numerous accounts in theRussian journals, some of them,fortunately, translated into English.These listed locality after locality, andthey gave great catalogues of speciesfound at each site. But it is very hard toget a real feel for a segment of the

history of life just from written accounts.We wondered also how accurate theaccounts were. If a Russian were to tryto understand the palaeontology ofEngland, or North America, or SouthAfrica, from old written summaries, heor she might receive quite the wrongimpression.

On the first field day, we drove backto the main road, back to Orenburg, andthen straight south towards the borderwith Kazakhstan, only a few kilometresaway. In heading south, we crossed theUral River in the middle of Orenburg,and thereby crossed from Europe intoAsia. We were heading for the Sol’Iletskdistrict. Here, in the banks of the Donguzand Berdyanka rivers, abundant remains

of fossil amphibians and reptiles weredug up in the 1940s and 1950s by IvanEfremov and his successors, includingOchev and Tchudinov, who were withus.

At one site we found dozens ofbones of the amphibian Chroniosuchus,not a temnospondyl, but a member of theother major Permian group, theanthracosaurs. These had narrowerskulls than the temnospondyls.Elsewhere, along the Donguz River,Ochev led us to site after site where hehad excavated great dicynodontskeletons using bulldozers in the 1950sand 1960s. The bulldozers were clearlyan important innovation, and theycertainly helped him to excavate huge

numbers of skeletons. Indeed, he seemsto have been so efficient at excavatingthat we didn’t find any at all. We wenton to visit some of the old copper minesin the Upper Permian, also south ofOrenburg; lumps of copper ore couldstill be picked up and the old workingswere still visible, but bones are nowextremely rare in the mine tailings.

The following day was a LowerTriassic day. We drove from the camp tothe graida, and headed westwards toSaraktash, some 15 kilometres away.During the drive, we had passed over asequence of Upper Permian rocks andinto the Triassic. We headed north forthe village of Petropavlovka. Here, onthe banks of an ancient oxbow lake,

formed in a bend of the Sakmara River,were a series of famous fossil sites,numbered clearly by the Russian authors,Petropavlovka I, II, III, IV, V and VI.Here we appreciated the value ofvisiting the sites. The Russian geologicalpapers give these site numbers, and listthe fossils from each, but there was nomap. We were lost until that point.

Tverdokhlebov had come equippedwith Soviet geological maps of thewhole area, and we were impressed andamazed to see them. They were ofexceptionally high quality, in someregards better than the equivalent Britishor American maps. Not only did theyshow the topography, roads and villages,as well as an overlay of the rock units

and geological structures, but they alsoshowed the high and low stands of riversand river vegetation. Each year, as theice and snow melt, the Ural, and all therivers coming off the Ural Mountains,flood. The date of flooding is fairlypredictable, as is the extent of the areaflooding and silt. The bushes and reedsalong the river banks then have seasonalbursts of growth. All this is shown onthe map.

We noticed that each map had anumber and that the word ‘Sekretna’ hadbeen heavily scored off with a thick pen.Might it be possible now for us to buymaps? Perestroika, glasnost, and so on?No. Although beautifully mapped andengraved, and representing dozens of

person-years’ work, only 100 copies ofeach map had ever been produced. Eachnumbered map was lodged andcatalogued in specific governmentoffices. Even Valentin Tverdokhlebov,who had produced a number of theOrenburg region maps himself, did nothave his own copies. He had to signthem out and return them. Could we copythem? No.

We visited all the Petropavlovkasites, and saw where the sparse faunahad been collected, some amphibiansand procolophonids, as at Tikhvinsk, butalso a larger predator, the archosaurGarjainia, perhaps 2 metres long andwith a deep-sided skull, clearly adaptedfor feeding on medium-sized amphibians

and reptiles. The Petropavlovka bedsare somewhat younger than those atTikhvinsk, representing the YarenskianGorizont, and the faunas had diversified.Instead of four or five species, therewere more like ten in the Yarenskian.Life was recovering after the end-Permian event, but it still had a long wayto go before the ecosystems could rivaltheir richness before the crisis.

The Permo-Triassicboundary

After a rest day, and an extremelydrunken evening party with the local

farmers, we set out into the field again.Glenn and I, not used to downing vodkaby the pint, were rather white-faced aswe sat in the back of the van. We droveto a bluff near the camp, called Sambula,and followed through a sequence ofsediments. This was to be one of themost astonishing days of the entire trip.

The rock succession at Sambulashowed a major change half-way up.The basal part consisted of cycles ofmudstones and limestones that had beenlaid down on land, each beginning withan erosional layer and ending with asoil, perhaps representing repeatedclimatic changes from wet to dry overseveral thousand years. Here and therewe were shown sites where reptile

skeletons had been extracted in the past.At the top was a vast thickness ofconglomerate, a coarse unit composed ofgreat boulders of exotic rocks – greenishschists, pinkish granites, whitequartzites. These boulders had clearlycome from the core of the UralMountains, some 200–300 kilometresaway.

Valentin Tverdokhlebov had mappedout the distribution of the conglomeratebeds, and they formed vast alluvial fans,the outwash at the head of major riversystems. Clearly, there had been heavyrainfall over the proto-Urals 250 millionyears ago, and they must have beeneroding down rapidly. Great rivers inflood had transported massive boulders

down over the plain. At the top wasnothing: the softer rocks above had wornaway much later to form the modernlandscape. Following across country,though, the overlying sandstones andmudstones could be found.

‘Kak Gorizont?’, I asked.‘Eta granitsa Permo-Trias’ replied

Tverdokhlebov nonchalantly.This took a minute or two to sink in.‘The Permo-Triassic boundary?’, I

repeated to Glenn. ‘He says this is thePermo-Triassic boundary. How does heknow? What’s the evidence?’ Glennquestioned Valentin Tverdokhlebov indetail in his more fluent Russian. If thiswas truly the Permo-Triassic boundary,

it was important to understand howaccurately it had been dated, whatenvironmental changes had happened inthe switch from wet-dry cycles tomassive alluvial fans, and how the plantand animal life had been affected.

Back in camp that night, as themosquitoes feasted on us and thedragonflies soared and swooped afterthem, Glenn and I had a great deal to talkabout. If this really was the Permo-Triassic boundary, we had to find clear-cut evidence to justify that claim. Thenwe could actually follow inch by inchwhat had happened all those years ago.But we had to leave the Sakmara the nextday to begin the long journey home. Wemust come back.

Across the boundary

Our 1994 trip had just been a shorttaster. Glenn and I organized a moreserious expedition for July and August1995, in collaboration withTverdokhlebov and Ochev. We campedin the same place as before, but with aneven larger crew. Two of my students,Andy Newell, now employed by theBritish Geological Survey, and PatrickSpencer, an expert on small reptiles,went out in advance to join the Russians.Glenn and I arrived two weeks laterwith David Gower, another student, nowworking at the Natural History Museumin London, and Darren Partridge, a fossilenthusiast who works as a vet.

When we arrived in camp, we found13 residential tents and a large kitchentent containing a full-sized domesticcooker, operated from a gas cylinder,two cooks and two dogs. In all therewere 18 Russian geologists and others –most were there, no doubt, for areasonably-priced holiday by the river.We had agreed to pay for the costs of thecamp (the Saratov geologists had paidfor everything in 1994), and we handedover the sum of $5000 in roubles. I hadtaken the dollars to Russia in smalldenominations, new 1- and 5-dollarnotes, and changed them into roubles atthe airport. I had felt very vulnerable onthe train, travelling with a bundle of 18million roubles in a wad as thick as a

brick in a moneybelt round my neck.We soon settled into camp life,

rising early and eating a huge cookedbreakfast, usually consisting of cashaand gristle. Casha is a classic Russiancarbohydrate, something like buckwheat,harvested from low plants that growwell in the steppe soils, and tasting alittle like rice. The casha was alwaystopped with some meat and gravy. Thiswas served for breakfast, but generallyalso for lunch and dinner. After a hardday in the field, and the weather wasusually hot, we were happy to return tocamp each evening and swim in theriver.

We planned four weeks offieldwork, and hoped to be able to

document the sedimentology in detail towork out exactly how environments andfaunas had changed through the LatePermian and Early Triassic. Every daywe went out to a new site. ValentinTverdokhlebov had visited them allbefore, but we were keen to seeeverything ourselves. So he indulged ourdemands to see more and more of thePermo-Triassic rocks. We raced up anddown ravines making sedimentary logs,that is, recording bed-by-bed the natureof the sediments and any included fossilsor sedimentary structures. Sedimentarystructures can give clear guidance on theenvironments of deposition. Forexample, different kinds of preservedripples show the direction of water

movement, and may indicate whether thesands were deposited on a beach or in ariver. Mudcracks and rain prints ofcourse indicate exposure to the air.Impressions of ancient roots point to asoil. Channels small and large canindicate the type of river, whethermeandering or braided, and so on.

In teams of two we were able tocover the ground fast, logging hundredsof metres of sedimentary sequences,right through the Late Permian and theEarly Triassic, and frequently includingthe Permo-Triassic boundary. Oneperson would carry the tape measure andwould shout out the thickness, the rocktype and any sedimentary structures orfossils. The other wrote it all down.

The beauty of sedimentary logs isthat you can convert all themeasurements and hasty scribbles aboutrock types in your field notebook into asimple diagram back in the lab. Thediagram (Fig. 39) gives a summary ofsometimes hundreds of metres orkilometres of thickness of sediments in acartoon form that can be read directly.While you are on the ground, fightingthrough scrub and undergrowth up awinding ravine, the grand pattern mayseem obscure. Laid out on a large sheetof paper or on the computer screen, thebroader picture may become clear.

39 A typical sedimentary log,showing the succession of rocksdeposited by Late Permian andearliest Triassic rivers washingdown from the Urals. Thesuccession is divided into threecolumns, beginning at the bottomleft, and ending at the top right.The sediments show evidence formudflats low down (column A),passing up into sandy channels(column B), gravelly channels andthen a major gravelly channel(column C).

Megafans

One thing that was immediately obvious

in the field was the apparently suddenappearance of massive conglomerates aswe had seen at Sambula on a previousexpedition. This was not a local featureof just that site, but appearedeverywhere around the South Urals.Something dramatic had happened – alarge-scale change in sedimentationpatterns.

Andy Newell, our sedimentologist,worked through all the sedimentary logsand was able to establish the scale ofwhat had happened at the Permo-Triassic boundary in Russia.1 The rocksuccession begins with mudflats,homogeneous mudstones with rare, thinsandstones. The muds were deposited inshallow pools that were surrounded by

plants, and reptiles and amphibiansoccasionally wandered across, as shownby some fossilized footprints. Cracks ina number of the mud layers show that thepools dried up from time to time andimpressions of salt crystals are evidencethat some were at least partially saline –they perhaps flooded and dried up manytimes, concentrating the salts.

The mudflats are followed by asequence containing much moresandstone, representing sandydistributary systems. The sandstonesoccur as channels and as sand sheets,suggesting a variety of types of rivers,from well-defined flows in long-lastingchannels, to dramatic floods thatpresumably swept huge amounts of sand

out over a wide area, clothing the wholelandscape. This kind of flooding impliesa semi-arid climate, with occasionalheavy rainfall that erodes thesurrounding hills and washes masses ofsand over the whole area.

These channel and flood sands arefollowed by the third sedimentaryassociation, small gravelly channels.The channels are about 100 metres wide,and 1 or 2 metres deep, and they arecomposed of hand-sized pebbles ofquartzite, limestone and marble. Theserepresent wide, shallow riverstravelling at some speed, and carrying inrocks from 100 kilometres away, fromthe foothills of the Ural Mountains. Thepebbles are well rounded, confirming

that they have been rolled and tumbledsome distance by the rivers.

Then come the massiveconglomerates we first saw at Sambula.These are on quite a different scale fromanything that had gone before – hereagain something dramatic has clearlyhappened. Individual layers arerelatively thin, 1 to 3 metres deep, butthey may be 3 to 5 kilometres wide. Theincluded fragments are mainly pebbles,but boulders commonly occur. As atSambula, the pebbles are of quartzite,limestone and marble, as in the thinnerconglomerates below, and come fromthe western edge of the Ural Mountains.Added to this array of transported rocksare also boulders of chert and igneous

rocks whose source has been traced tothe core of the Ural Mountains, as muchas 900 kilometres to the east.

Valentin Tverdokhlebov has beenable to track most of the rock typesfound mixed randomly in the massiveconglomerates back to their originalsources in the heart of the UralMountains. Each of the conglomeratebodies clearly mixes material from anumber of source areas, and this provesthat there were several huge, fast-flowing torrents which hurtled down themountain sides, rolling and tumblinggreat boulders. As the torrents came tothe lower slopes, they merged into widerrivers, and the boulders of marble andchert and igneous rock became mingled

in enormous rivers that carried millionsof tonnes of rocks.

40 Map of the distribution ofearliest Triassic megafans washingoff the Ural Mountains to the right.River flows are indicated byarrows which pass into the fans,five of which are shown (A–E). Infront of the fans, to the left in thediagram, are finer sediments ofrivers and lakes.

By careful mapping, Tverdokhlebov2

was able to work out the exact shapes ofthese huge outwash systems, termedalluvial fans, or, more properly,megafans, since they are several ordersof magnitude larger than the gravel fansbelow. According to his maps (Fig. 40),each megafan had an area of some 300square kilometres, and extended 15

kilometres out from the Ural foothills.

Crisis

So why the huge change in sedimentationpattern? Newell suggests that themudstone to sandy distributary system tosmall gravelly fan succession in thelatest Permian is a linked series ofdeposits. All of them imply the samekind of processes, as sediment waswashed into the South Urals area fromthe surrounding hills at different rates,which depended partly on rainfall andpartly on the subsidence rates.

Some 60 million years before the

end of the Permian, two great tectonicplates representing the precursors ofwhat are now Europe and Asia met, andslowly pushed against each other,raising the Ural Mountains as they wereforced together. A similar process isgoing on today, with India thrustingagainst the rest of Asia – its statelynorthwards progression is forcing theHimalaya ever higher (but only bycentimetres per century).

In the Late Carboniferous, as theUral Mountains began to rise, thesurrounding landscape sank. At first,seas covered much of that part of Russia,and thick marine sediments accumulated.As these filled up the basins, land waseventually formed in the Late Permian.

During the Permian, subsidence ratesdecreased as the Ural Mountainsstabilized, and the old plate movementsfizzled out. The change from mudstonesto sandstones to small gravelly channelsin the latest Permian represents the lastphases of filling of the basin, a shift fromlow-lying areas with ponds to higherlevels more in line with the foothills ofthe Urals.

Then come the massiveconglomerates. The easiest explanationmight be that there was a suddenincrease in rainfall. However, all theevidence of the sediments points to areduction in rainfall, and a spread of dryconditions. More likely is that there wasa huge reduction in vegetation cover at

this time, exactly at the Permo-Triassicboundary. The loss of plants, as RogerSmith, Peter Ward and colleagues hadsuggested in South Africa, leads tomassive runoff. First, all the soils thathad been held in place by plant rootswould go. Any rainfall would then havea devastating effect on the strippedhillsides; the bare rock would rapidly becut into deep ravines, and fast-flowingtorrents carrying boulders would smashfurther rocks out of place, until the riverbecame a moving liquefied flow of rockdebris.

So the two teams, Smith, Ward andothers in South Africa, and our group inRussia, had seen exactly the sameprocesses in action, in two widely

separated parts of the globe. Evidentlysomething had happened at the Permo-Triassic boundary to strip the world ofvegetation, and this is borne out by thesubsequent ‘coal gap’ and the ‘fungalspike’, as we have seen. In looking foran explanation of the end-Permian crisis,as we shall in the next chapter, it isimportant to bear these facts in mind.

The reptiles take a dive

Skeletons of amphibians and reptiles arefound throughout the Late Permian rocksequence around the South Urals. Theydisappear abruptly at the top of themudstone-sandstone-small gravelly fan

succession, and then reappear above themegafan deposits. In one sense, thisdisappearance is a local phenomenon,because you would not expect to findrelatively delicate bones in acatastrophic flood deposit. But, abovethe megafan beds, the amphibians andreptiles that survived the crisis into theearliest Triassic in Russia are a poorassemblage, as we saw earlier in thischapter.

The complex latest Permiancommunities of small procolophonids,larger herbivorous dicynodonts andpareiasaurs, preyed upon variously bytherocephalians, cynodonts and thesabre-toothed gorgonopsians likeInostrancevia, collapsed. The earliest

Triassic assemblage, the so-calledLower Vetluga (Vokhmian) Community,Efremov’s old Zone V, was reduced toLystrosaurus as the sole, reasonablysized herbivore, one species ofprocolophonid, and some raretherocephalians and diapsids asinsectivores and carnivores that preyedmainly on the smaller reptiles.

This information had been reviewedin English by Efremov and colleagues,and by Everett C. Olson, a distinguishedAmerican vertebrate palaeontologistwho had visited Russia several times inthe 1950s and 1960s. However, thosepublications3 were now very out of date,and much had been found since. Theproblem was partly one of language –

Glenn could speak Russian well, but Ilaboured to make myself understood, andfound reading Russian a very slowprocess. So we hatched the plan ofpublishing a book, in English, that wouldsummarize everything that was known ofthe Russian Permo-Triassic reptiles.After much anguish, and many heateddebates, the book4 finally appeared inDecember 2000, nearly 700 pages longand containing 30 articles by variouscombinations of 34 palaeontologists,Russian and western. That was onething. But we still did not have a clearpicture of how the faunas were changing.

As we raced around the South Uralswe saw site after site, each of which hadproduced some skeletons of amphibians

and reptiles. We found the mass of newinformation hard to take in, and asked ifanyone had actually done a census of therise and fall of the different animalgroups through time. Indeed, Valentintold us, he had compiled a hugecatalogue of all the sites. When wereturned to Saratov with him after theexpedition, an interesting journey of 700kilometres in the front of a huge Gaz 66truck, Valentin showed us his cardindex. In it, 400 or so sites are listed,each with a determination of itsgeological age and a list of the fossilsthat had been found there.

A big job for the future will be towork with our Russian colleagues totranslate the card index, and plot the

information it contains about the rise andfall of reptiles across the Permo-Triassic boundary. A preliminary scanof the information suggests that thepattern is just the same as in SouthAfrica – a catastrophic drop in diversity,followed by a long and slow recovery ofecosystems.

The final solution

So what caused this, the biggest of allmass extinctions? The nature of whathappened is now much clearer than ithad been, say, in 1990. Careful datinghas shown that the event took place 251million years ago, and that species

losses were anything from 90 to 95%.This was no local phenomenon, since ithas been detected in rocks from China toSpitsbergen, from Greenland to SouthAfrica, from Russia to Australia. Inevery case, whether looking at events onland or in the sea, the rate of speciesloss seems to have been similarly huge.There were no safe refuges, nowhere tohide.

There is also no evidence forselectivity, except that the survivorstended to be widespread species. Butwith such a tiny survival rate – 10% orless of species made it through – it isclear that plants and animals were beingwiped out with almost no regard to theiradaptations. Certainly on land the large

animals all disappeared, but this couldbe explained as part of a chanceprocess. If there are 100 species, ofwhich 10 are large, a 95% loss ofspecies is likely to kill all the largeanimals. Add to that the probability that,as today, large animals are rarer thansmall animals (i.e. smaller populationsizes), then it is easy to see why all therhinos and elephants might disappear,but a few rats and squirrels mightsurvive. It would be wrong, though, tosay that this proves that individual ratsand squirrels are better adapted tosurvive crises. It is perhaps only theirgreater abundance that protects them as aspecies.

We have some hints of the

environmental changes too. In the sea,the rocks show an increase in anoxia.Many of the surviving marine creaturesseem to have been peculiarly adapted toliving in such conditions of low oxygen.There are hints too of low productivity,meaning a lack of organic matter in foodchains, so many of the surviving specieswere presumably able to survive on verylittle food.

On land, as the recent studies inSouth Africa and Russia have shown, theend of the Permian is marked by asudden change in sedimentation, withmegafans composed of huge boulders. Inneither case can this be explained by adramatic increase in rainfall. Indeed, theevidence suggests increased aridity, so

the dramatically heightened levels oferosion and runoff can best be explainedby a sudden loss of vegetation and soils,perhaps worldwide. Soils show thatclimates also became warmer.

So what caused the crash? The eventmust have been sudden. It must havereduced oxygen levels, increasedtemperatures and reduced rainfall, all ona worldwide scale. It must have had theability to push all of life virtually to thebrink. The tentacles of the killing agentreached into shallow seas and into thedeepest oceans. On land, they penetratedlowland basins and mountainousregions, rivers and lakes. What kind ofcrisis could have been so profound thatit killed reptiles on land, and

brachiopods and corals on the sea floor?It cleared the Earth of vegetation, even iffor a short time. This is more profoundthan any of the puny threats that humanshave devised so far, whether nuclearbombs or mass forest clearance. Aglobal rise in temperature of half adegree in a century as a result ofindustrial pollution and global warming?That fades into insignificance beside thecrisis 251 million years ago.

11

WHAT CAUSED THEBIGGEST CATASTROPHE

OF ALL TIME?

I remember as a child reading ArthurMe e ’s Children’s Encyclopaedia inwhich the kings and queens of Englandwere catalogued. Henry VIII was badbecause he had six wives. But hisdaughter, Elizabeth I, was good: she hadhad an unhappy childhood – her mother

was beheaded and she was declared abastard by Parliament – and as Queenshe suffered smallpox, plots to kill herand the Spanish Armada. Henry diedunhappy, while Elizabeth had a long andglorious reign.

This sort of account may makehistory easier for children to understand,and the weaving of moral tales used tobe a favourite way to encourage childrento become good citizens. But thisapproach is now much criticized.Historians are taught not to judge thepast by present-day values. CharlesDarwin said that certain human raceswere superior to others. By modernstandards, he clearly was a racist. Butby the standards of England in the 1850s,

he was probably considerably less racistthan many other people.

This interpretation of history iscalled ‘Whig’, a term invented byHerbert Butterfield, in a little book ofthat title1 published in 1931. Butterfieldwas concerned that history was oftenwritten as if there were a preordainedend to the story, and that moral lessonswere to be learned. I had alwayswondered why Butterfield chose theterm ‘Whig’. I knew that the Whigs andTories were the main political parties inBritain in the earlier part of thenineteenth century, and that the Whigsbecame the Liberals after about 1860. Itseems that Butterfield’s term derivedfrom the practice of history writing in

the 1830s, when Thomas, LordMacaulay (1800–59), who waspolitically a Whig, adopted thedirectional and moralistic approach.Benjamin Disraeli (1804–81), the greatTory prime minister and author,constructed an alternative view ofEnglish history and parodied the Whigapproach. (As for the strange term‘Whig’, it seems to have been derivedfrom a derogatory term applied to agroup of Scottish presbyterians duringthe seventeenth century who opposed theCatholic succession to the throne. In full,the term was originally ‘whiggamore’,coming perhaps from the old Scots, to‘whig’, or urge along, and ‘more’, amare.)

As for the history of kings andqueens, so too for science. Looking backto the state of understanding of the end-Permian crisis in 1970, or 1980, or even1990, we can easily point to errors ofjudgment. Why were palaeontologistsand geologists then so blind to the truth?(Can we be sure we have the truth now?)They were looking at the biggest crisisin the history of life and of the Earth andthey didn’t see it. They must have beendeluded, or incompetent, or both. Not so.It is hard for scientists, just as foranyone else, to throw off everything theyhave been taught. And when theevidence is somewhat intangible, it isunderstandable if scientists err on theside of caution.

Pervasive conservatism

As we saw earlier, it was commonplacein the 1960s to deny that anything of notehad happened at the end of the Permian.Lone voices, like Otto Schindewolf inGermany, who claimed that there hadreally been an astonishing andcatastrophic crisis, were crying in thewilderness. By 1970, however, throughthe work of Norman Newell and JimValentine, who had carefully cataloguedthe rise and fall of different groups ofanimals in the sea, the end-Permian massextinction stood out in ever sharperrelief. John Phillips’ intuition of the1840s was borne out, and the hiatusbetween the life of the Palaeozoic and

the Mesozoic could not be explainedaway by some failure of preservation.

Geologists did not, however, thenassume that the great dying-off had beencatastrophic. The evidence just wasn’tthere. Indeed, although we now knowthat Schindewolf was broadly correct,he had based his judgments on reallyrather inadequate evidence. Dating of thePermo-Triassic boundary in the 1960swas subject to considerably more errorthan it is now, and detailed, inch-by-inchstudies of high-quality Permo-Triassicboundary sections, such as those inChina, Italy and Greenland, had yet to bedone. Indeed, a common assumption thenwas that the boundary was marked by agap in deposition, and nothing much

could be got out of it anyway.In 1967, Frank Rhodes, a British

palaeontologist who had becomePresident of Cornell University, gave thebest explanation of the end-Permianevent then possible,2 that it had beencaused by ‘the multiple interactions of awide variety of physical and biologicalfactors’. Even in 1993, Doug Erwin, inhis masterly review,3 was unable to gomuch further: ‘I believe that theextinction cannot be traced to a singlecause, but rather a multitude of eventsoccurring together’. There had beenmuch to-ing and fro-ing before thesecomments were written, but no hint of aconsensus model that fitted all theevidence.

Continental fusion

Most popular around 1970 was the viewthat continental fusion had caused along-term extinction event that lastedperhaps 20 million years. NormanNewell, as early as 1952, had noticedthat there had been a major marineregression4 during the Late Permian. Heargued that a general fall in sea levelwould have led to a reduction in oceanvolumes, and hence a fall in availablehabitats on the sea bed and in the watercolumn above. So, as sea levels fell, heargued, species of marine animals wouldhave died out progressively, until thenumbers were much reduced. Of course,at that time the true magnitude of the end-

Permian extinction was perhaps notappreciated.

Newell at first could not explainwhy sea levels had fallen, but platetectonics provided the answer. AlthoughAlfred Wegener (1880–1930), a Germanmeteorologist, had proposed continentaldrift in 1915 – and some of his bestevidence came from Permo-Triassicplants and animals – many geologistsrejected the whole concept. They arguedthat the Earth’s crust was obviouslysolid and not mobile, and in any casethere was no mechanism to drive theplates around.

The discovery of the function of mid-ocean ridges and the existence ofsymmetrical rock sequences on the

seafloor around 1960 proved Wegenerhad been right. Running in a strip downthe middle of the Atlantic is a zone ofactive upwelling of magma, the Mid-Atlantic Ridge. As magma wells up, itforms fresh rock on either side of theridge, and slowly the plates beneath theeastern and western halves of the oceanmove sideways. This was proved by thediscovery of symmetrical, datablesuccessions on the floors of both easternand western plates, which showed ahistory of slow opening of the AtlanticOcean over some 200 million years. Asthe Atlantic plates move east and west,the continental plates beneath NorthAmerica and South America are alsomoving west, while those beneath

Europe and Africa move east. Since theEarth is of fixed size, plates are alsobeing consumed or crushed in other partsof the world.

It is now known that the tectonicplates move at the same rate as thegrowth of human hair. That is about 5centimetres per year. This stately paceof the continental and oceanic plates hasbeen established by back-calculationsfrom the present position of thecontinents to their former locations,confirmed by measurements bounced bylaser off the moon. As the opening NorthAtlantic pushes North America west, itconverges on Russia. The Pacific plate,carrying San Francisco and Los Angeleson its eastern edge, is heading northwest,

towards Japan. India and Australiacontinue their northwards migration.

With the acceptance of platetectonics in the 1960s, Wegener’s modelfor the ancient supercontinent Pangaeawas also clearly correct. During thePermian period, isolated continentsbegan to fuse together. As we have seen,Europe and Asia had already combinedin the Mid Carboniferous with the upliftof the Ural Mountains. Laurasia, thenorthern continents, fused withGondwana, and this was largelycompleted during the Late Permian. So,Jim Valentine and Eldridge Mooresargued,5 as the continents fused, ancientseaways closed off. Worldwide, it wascalculated, sea levels fell by 200 metres

or more. How could this causeextinctions?

Valentine and Moores argued fortwo elements to the killing model. First,the global fall in sea level reduced theoverall number of habitats on the seafloor and in the water column, as Newellhad argued. But there had been othermajor changes in sea level during thepast 600 million years, and many of themwere not associated with huge levels ofextinction. What was special about theLate Permian episode was the secondkilling mechanism, the reduction inendemism. Endemism is a biologicalterm referring to the restrictedgeographic distribution of certainspecies. So, as the continents fused

together, animals and plants on landcould clearly move freely from place toplace. This would mean that endemic, orlocal, species would either becomecosmopolitan, or they would die out,overwhelmed by invading forms fromelsewhere. In the seas too, as isolatedseas and basins disappeared, all theirendemic species would go too. Imagine,Valentine and Moores were suggesting,that the Mediterranean, the Caribbeanand the China seas, and all the seasaround Australia and the southeast Asianislands, were to disappear. All thediversity of their endemic species ofcorals, molluscs and fishes woulddisappear into an homogeneous globalfauna. There would be a global crash in

diversity.This theory seemed initially

appealing, and it did seem to account forextinctions on land and sea. It was alsosafe: no need for cosmic rays orimpacts. However, there are threereasons why it doesn’t explain the end-Permian crisis. First, some censusingexperiments based on moderndistributions of marine organismsshowed that the loss of a few inland seascould never account for a loss of 90–95% of species. Second, the assumptionof a major regression during the LatePermian has been questioned. Somegeologists deny that there was anyregression at all, while others6 argue thatit was at most a minor affair, since there

is little evidence for the emergence ofland or for large-scale erosion. Indeed,sea level was actually rising rapidly atthe time of the extinction.

The third reason to reject thePangaea/regression model is that it is notfast enough. In 1973, when it wasproposed, everyone was convinced thatthe extinction had lasted for 20 or 30million years, so a slow process lastingfor much of the Permian was required.Now that the event is thought to havebeen much more rapid, regression andcontinental fusion just will not do.

But how can one go from apostulated long-term Late Permianextinction lasting for 10 million years ormore to an instant event? If all the

extinctions are now concentrated right atthe end of the Permian, what happened tothe earlier ones? Have they all beensubsumed into the terminal Permian insome kind of large-scale correction of amega-Signor-Lipps effect (artificialbackward smearing of apparent times ofextinction). Perhaps there was more thanone event?

How many events?

As we saw in Chapters 7 and 8, the end-Permian mass extinction is dated at 251million years ago, and close studies ofthe Meishan section in China haveshown that 95% of species died out at

that time, or within the following800,000 years. There are still debatesabout the precise timing of the event, andwhether it occurred at the same timeeverywhere, and at the same time onland and in the sea.

Actually there seem to have beenthree pulses of extinction across thePermo-Triassic boundary, separated byconsiderable spans of time: one tookplace 5–10 million years before thePermo-Triassic boundary, and another5–10 million years after. First hints ofthese extra peaks of extinction camefrom studies of continental vertebrates.At a time when synoptic plots of marinediversity change just showed a gradualdecline through the Late Permian, the

plot7 of amphibian and reptilediversification highlighted threeextinction peaks, one at the end of theCapitanian Stage, one at the end ofChanghsingian Stage, and one during theOlenekian Stage. The ChanghsingianStage is the last time division of thePermian, named after the Chinese rockunits of the Meishan section. The end ofthe Capitanian Stage, lower down in theLate Permian, named from classicsections of the Capitan reef in Texas, isdated variously at 255 or 260 millionyears ago, while the end of theOlenekian Stage, named after the RiverOlenek in the South Urals,corresponding to the end of the EarlyTriassic, is dated as 240 to 245 million

years ago. So was the extinction a triplewhammy?

The end-Capitanian mass extinctionsaw the loss of up to 58% of marinegenera, according to calculations bySteven Stanley of Johns HopkinsUniversity and X. Yang of NanjingUniversity: less than at the end-Permianevent itself, but more than during the KTevent. Many families and generacertainly disappeared, with majorreductions in the diversity ofammonoids, rugose corals, bryozoans,fusuline foraminifera and articulatebrachiopods. Of major marine groups,only the blastoids disappeared. Detailedstudies of the fates of 753 species ofcorals through the Late Permian in China

have shown that 76% of families, 78%of genera and 82% of species died out atthe end of the Capitanian. Theextinctions happened especially inTethyan, mid-latitude regions; there isvery little evidence for the event innorthern waters.

On land there were also notableextinctions at about the end of theCapitanian Stage. The dinocephalianreptiles, with nearly 30 genera of plant-and meat-eaters present in theTapinocephalus Zone of the KarooBasin, disappeared. So too did somebasal dicynodont relatives and someflesh-eating synapsid groups, andessentially the same patterns of loss areseen in Russia.

The Olenekian event, just before theend of the Early Triassic, is oftenforgotten. After all, life was recoveringduring the Triassic. But there wasindeed a third blow, five million yearsafter the end-Permian event. There werehigh rates of extinction amongammonoids, and perhaps amongbivalves. In addition, many groups ofamphibians disappeared. But moredetailed study of marine sections isrequired. On land, the Karoo Basincannot offer much help since it is notclear that the sediments of the BeaufortGroup span this interval, but the Russiansuccessions show a major loss ofamphibian species.

Despite the interest of these extra

extinction events, bracketing the key,end-Permian mass extinction some fivemillion years before and after, they arerather poorly defined at present, and arecertainly smaller in magnitude than thebig event. If it turns out now that the end-Permian mass extinction was ageologically rapid event, is this a reasonto look for extraterrestrial mechanisms?Some people think so.

Cosmic rays and impacts

The impact scenario for the KT event( s e e Chapter 5) immediately ledresearchers to seek evidence for impactat the time of other mass extinctions.

Intense efforts have been made to findany hint, however feeble, that the Earthwas struck by a giant asteroid or comet251 million years ago, and a recentstrong contender has been put forward,as we saw at the start of the book. Ofcourse, extraterrestrial catastrophe hadbeen posited before.

In 1954, Otto Schindewolf 8 hadsuggested that a burst of cosmic rays hadcaused the end-Permian mass extinction.He felt driven to this conclusion since hewas convinced from his studies of thePermo-Triassic boundary sections in theSalt Range of Pakistan that the boundarywas abrupt, and that life had been wipedout rather rapidly. Schindewolf, as wesaw (Chapter 4), was ridiculed for his

far-out ideas, and certainly noindependent evidence has been found forcosmic rays. What of impact?

There was great excitement in 1984when an iridium anomaly wasannounced at the Permo-Triassicboundary in China.9 This came only fouryears after Luis Alvarez and his teamhad published their epochal model forKT impact, and enhancement in the rareelement iridium was then the keyevidence for impact. Two teams ofChinese researchers reported 8 parts perbillion and 2 parts per billion of iridiumin the boundary layer, 10 to 40 timesnormal background levels. In 1986,however, two American teamsreanalysed the Chinese boundary clays

and found barely detectable levels ofiridium; if anything their values werelower than might be expected in normalsediments.

Some other possible hints of impacthave also been rejected. In the Chinesesections, some iron-rich spherules werereported, but these seem to have beenmore typical of a volcanic origin than animpact.10 But there was no shockedquartz, such a potent indicator of impactat the KT boundary.

Then the news broke. GregRetallack, well-known expert on ancientsoils from the University of Oregon,showed some pictures at the 1995 annualmeeting of the Geological Society ofAmerica that were claimed to be

shocked quartz from the Permo-Triassicboundary of Australia and Antarctica.As Richard Kerr, science journalist,reported,11

the hallways buzzed with palaeontologists andgeologists exchanging opinions on Retallack’sphotos, which purportedly showed faint bandsof glass-filled fractures within the grains.Retallack thinks the fractures formed in theshock of a massive impact and notes thatsimilar grains have been linked to theCretaceous-Tertiary extinction. The hallwaybuzz was more cautious …

Retallack’s critics, including PhilippeClaeys, then at the Humboldt Museum inBerlin, noted that his grains had beenoverprinted with later stress damage,and that most of them showed only one

set of planar deformation features(PDFs). Shocked quartz from KTsections typically shows three or moresets of such linear features. Kerr goeson,

Retallack counters that Claeys and othershaven’t seen all there is to see. Under themicroscope, where the full depth of a quartzgrain can be viewed by changing the depth offocus, all the grains can be seen to have at leastthree sets of PDFs, he says; one has seven.

The full account was published threeyears later, and here Retallack andcolleagues showed illustrations of theshocked quartz grains, some apparentlyshowing two sets of PDFs. The rocksuccessions in Australia and Antarctica

also showed iridium spikes at thePermo-Triassic boundary. The authorsaccepted, however, that the iridiumlevels were lower than in most KTsections, and that the PDFs in theshocked quartz were not so clear, so theevidence is still equivocal.

Buckyballs and impact

Further excitement broke out in February2001, as we saw in the Prologue, withthe announcement of a meteor impact.The evidence for this came in the formof extraterrestrial noble gases infullerenes at the Permo-Triassicboundary in China, Japan and Hungary.

Fullerenes are large molecules ofcarbon, composed of 60 to 200 carbonatoms arranged as regular hexagonsaround a hollow ball. They are oftencalled buckyballs, after RichardBuckminster Fuller (1895–1983),inventor of the geodesic dome, whichtheir natural structure resembles.Fullerenes can form in meteorites, inforest fires, even within the massspectrometers that are used in studyingthem.

In the 2001 Science paper, LuannBecker of the University ofWashington,12 and colleagues, reportedthat they had found helium and argon,two of the noble gases, trapped insidethe ‘cage’ formed by the carbon atoms.

Their samples came from Hungary,China (from the Meishan section) andJapan, from precisely the Permo-Triassic boundary. The samples fromHungary did not produce any particularsignal, but the Chinese and Japanesesamples gave strong evidence forimpact. When Becker and her teamanalysed the helium and argon fromthese sites, they found that it waschemically the same as helium and argonderived from meteorites, and it waspresent in relatively large quantities. So,they argued, the Permo-Triassicboundary fullerenes must have comefrom the impact of a meteorite. Theresults were supported and criticized inequal measure by other geologists and

geochemists when the publication cameout.

Nothing daunted by such uncertainty,the press went mad. Nigel Hawkes ofThe Times of London was judiciouslycautious, and simply reported the results,but his piece was accompanied by alurid colour illustration showing amassive asteroid hitting the Earth.Michael Hanlon of the Daily Mail wasless circumspect:

There was no warning. One minute, the Earthwas quiet, just as it had been for countlessmillennia.… Then it came. Out of the sky, witha roar that must have sounded like theslamming of the doors of Hell itself, an objectthat would change the course of life for ever onour planet. When it hit, the cataclysm it

unleashed was beyond imagining. The groundshook with the force of a million earthquakes,and a crater hundreds of miles across waspunched in the crust.

So there we have it. The case is proved.Well, actually not yet.

Laboratory procedures anddating problems

The critics, inevitably, were out in forcefrom the first announcement by theBecker team. Several geochemistsraised doubts about the laboratoryprocedures that had been used: after all,the analyses were subtle and involved

minute quantities of chemicals. Othersurged caution: ‘This is tricky stuff, anduntil it is confirmed there is little reasonto get too excited’, said Doug Erwinsoon after the Science paper appeared.Becker responded, ‘I have a feelingwe’re either going to go down in flames,or we’re going to be heroes.’

The measured, published criticismsfollowed in Science seven months later,in September 2001. First, K. A. Farleyand S. Mukhopadhyay from Caltech, inPasadena, California, stated that theyhad reanalysed samples from exactly thesame sites and using exactly the samelaboratory procedures, and yet they hadfailed to replicate the results of Beckerand her team. The Caltech scientists

studied material from the boundarylayers in the classic Meishan section,from bed 25, the boundary unit (seeChapter 7). They dried, weighed andcrushed the sediment samples, heatedthem to 1400°C to fuse them into a glass,and heated the sample further to 1800°Cto drive off helium. They found onlyminute quantities of helium, about 100ththe amounts reported by Becker, exactlyas expected in normal rocks that had notbeen affected by impact. The conclusionof Farley and Mukhopadhyay was:

We thus find no evidence for the impact-derived [helium] reported by Becker et al. Ouranalytical procedure for [helium] is as sensitiveand precise … as that used by Becker et al., sothe discrepancy between our results and theirs

is probably not analytical in origin…. Withoutconfirmation of fullerene-hosted [helium] inBed 25, both the occurrence of anextraterrestrial impact and the cause of themass extinction … must remain openquestions.

So, Luann Becker and her colleagueshad not found an enhancement of heliumin their samples from Hungary, andFarley and Mukhopadhyay had failed toconfirm high levels in the classicChinese samples. So what of the thirdlocality, Japan? Becker had obtainedsamples from the Sasayama section insouthwest Japan, but in a second criticalcomment in Science, Yukio Isozaki fromthe University of Tokyo argued that thePermo-Triassic boundary is missing in

this section. The samples analysed byBecker came from at least 80centimetres below the boundary, and henoted major geological disturbance ofthe rock succession. So, Isozakiconcluded, the Japanese samples are‘significantly older than Meishan Bed25,’ and cannot be related to the end-Permian event, whether it was an impactor not.

These two critiques would appeardevastating. However, Luann Beckerand her colleague Robert Poreda,responded robustly. The Caltech teamhad measured the wrong sample, itseems. Becker and Poreda had focusedon a carbon-rich layer at the base ofMeishan bed 25, and it was in this

horizon, which corresponds to the exacttime of maximum extinction intensity,that they found the fullerenes and theenhanced levels of extraterrestrialhelium. Farley and Mukhopadhyay hadsent Becker and Poreda some of thesample they had analysed, and it turnedout to consist mainly of sand grains, sowas obviously from a fractionally higherlevel within bed 25 from Meishan.

Meanwhile, Becker and Poreda hadreanalysed the materials from bed 25that they had included in their firstpaper, and they also analysed somereplicate samples from the same horizon,but collected at a different time. In allcases, they confirm high values ofhelium, 100 times or more the normal,

‘background’, levels. This aspect of thework will run and run, as yet moresamples are checked and rechecked indifferent laboratories. The work iscomplex, and requires careful weighing,drying and heating, as well as analyticalinstruments of astonishing precision.One sneeze or snuffle by the technicianat the wrong moment, a poorlycalibrated weighing machine or an offday for the mass spectrometer, and thereadings will be meaningless.

Isozaki’s comments on the dating ofthe Sasayama section in Japan wererather devastating, and Becker andPoreda had to admit that ‘the [Permo-Triassic boundary] cannot be preciselydefined in any of the Japanese sections

because of poor stratigraphic control’.They maintain, however, that since noone can prove their samples do not comeprecisely from the boundary bed, theywill assume that they do. And, theyconclude, defiantly:

Based on these new results, it would appear thatan impact event of global proportions remainsthe best explanation for the most severe bioticcrisis in the history of life on Earth.

And indeed, they seem to have somebacking from an unexpected source.Also in September 2001 Kunio Kaiho ofTohoku University in Japan and hiscolleagues reported sediment grains thatsupposedly show evidence ofcompression by impact, as well as

geochemical shifts that they interpret asindicating the impact of a huge asteroid.However, their data are far fromconclusive, and will doubtless becriticized. It is impossible to judgewhich way the debate will go in the nextweeks or months.

For the moment, however, I shalladopt here a conservative stance, andassume that the end-Permian crisis wasnot caused by an impact. The shockedquartz, and the fullerene and heliumevidence may be hotly debated, but thereis some other evidence concerning theend of the Permian that is much lessdebated – and it turns out to be just asdramatic and compelling.

The Siberian Traps

At the end of the Permian, 251 millionyears ago, giant volcanic eruptions tookplace in Siberia, spewing out some 2–3million cubic kilometres of basalt lava,covering 3.9 million square kilometresof eastern Russia to a depth of 400 to3000 metres. This region, known as theSiberian Traps, is equivalent to the areaof the European Community. Many nowthink that these massive eruptions,confined to a time-span of less than onemillion years in all, caused the end-Permian crisis.

The suggestion was first made in the1980s. Russian geologists had exploredthe Siberian Traps long before then, but

were not sure of their age. Over the hugearea of their distribution, the volcanicrocks lay variously on top of sedimentsthat could be dated from Silurian toPermian by their included fossils. Sothat meant they must at least be youngerthan the Permian rocks they covered.

The Siberian Traps consist of basalt,a dark-coloured igneous rock. Basalt iscomposed of plagioclase feldspar andpyroxene, with smaller quantities ofolivine and magnetite or titaniferousiron. The iron and magnesium in basaltgive it its black colour. The name isderived from the Latin basaltes, meaninga touchstone – a black stone that wasused by chemists to test silver or gold bythe colour of the streak they left. Basalts

form at temperatures between 1100 and1200°C – higher than the temperature offormation of granite, for example.

Basalt does not always come out of‘textbook’ conical volcanoes. Classicvolcanoes do produce basalt, as onHawaii and in Italy for example, but alsothe lighter-coloured andesite andrhyolite lavas, which are ejected inrapid and sometimes explosiveeruptions. In the case of flood basalts,the lava is ejected, seemingly much moresluggishly, from long fissures in theground. A major basalt flood eruptionoccurred in southern Iceland in 1783,when lava was erupted from the Lakifissure, 25 kilometres long. Icelandfamously lies athwart the Mid-Atlantic

ridge, and fresh ocean floor is createdby means of the eruptions; but here theocean floor forms land. The Lakieruption lasted from June to Novemberin 1783, and during those six months,about 150 cubic kilometres of lava wereerupted. The lava flowed widely,covering an area of 600 squarekilometres to an average depth of 250metres. Flood basalts typically formmany layers, and may build up overthousands of years to considerablethicknesses. They produce acharacteristic kind of landscape calledtrap scenery, as the different layers,through time, erode back, producing akind of layered, stepped appearance tothe hills. The word ‘trap’ comes from

the old Swedish word trapp, meaning astaircase.

Ancient flood basalts cover hugeareas of the Earth’s crust, on allcontinents, and many of them have beentied to former mass extinctions. Themost notable link has been madebetween the Deccan Traps in India andthe KT extinction event, as wasmentioned in Chapter 5. Early efforts atdating the Siberian Traps produced ahuge array of dates, from 280 to 160million years ago, with a particularcluster of dates between 260 and 230million years ago. According to theseranges, geologists in 1990 could onlysay that the basalts might be anythingfrom Early Permian to Late Jurassic in

age, but most likely they spanned thePermo-Triassic boundary. A wide agerange was not unexpected, since withthicknesses of as much as 3 kilometres inplaces, it could be postulated that theSiberian Traps had actually beenerupted sporadically over a period oftens of millions of years. Similar broadage ranges were associated with theDeccan Traps, and other thick floodbasalt successions.

In 1990, then, geologists weredivided about the significance of theflood basalts. Some enthusiasts, mostnotably Vincent Courtillot of theUniversity of Paris, argued that themajor flood basalts coincided with massextinctions. The basalts had destroyed

life, not so much by flowing over plantsand animals and killing them, as by theexpulsion of gases which altered theatmosphere and climate in catastrophicways. Other enthusiasts, of course, wereat that time also trying to explain allmass extinctions by asteroid impacts.

The vagueness of the radiometricdates did not help Courtillot’s case. Solong as the mass extinctions seemed tohave been geologically rapid but thebasalt accumulations spanned tens ofmillions of years, a link could not bemade. In 1988, Courtillot claimed thatthe Deccan Traps had been erupted inless than half a million years, whileother dating specialists disputed hismeasurements, and claimed that a longer

time-span was involved.13

Dating was all. The geologists couldmake out as many as 45 separate layerswithin the thickest sections of theSiberian Traps, in the northwest of theirdistribution, near Noril’sk. It wouldclearly be important to date the bottomand top of the pile to assess the entiretime-span of the eruptions. New methodsof more precise radiometric datingbecame available in the 1990s. In oneearly effort, Ajoy Baksi and EdwardFarrar14 seemed to discount any role forthe Siberian Traps in the end-Permianmass extinction. Their dates, using a newargon-argon dating method, spannedfrom 238 to 230 million years ago,corresponding to the Mid Triassic, and

long after the end-Permian event. Thisgave Courtillot and his supporters pausefor thought.

But not for long. Redating by otherresearch groups, using a variety ofradiometric methods, yielded datesexactly on the boundary, and the rangefrom bottom to top of the lava pile wasabout 600,000 years. In 1993, DougErwin, in his overview of the Permo-Triassic crisis, considered all theevidence and identified a role for theSiberian Traps eruptions, but only arelatively minor one, as one of severalsources of carbon dioxide. Indeed, he15

compared the end-Permian event to theplot of Agatha Christie’s novel Murderon the Orient Express, in which the

victim,

a singularly loathsome man, is found murderedin his compartment on the Orient Express withtwelve knife wounds. As Hercule Poirot listensto the stories of each passenger on the coach,he finds they eliminate one another as suspects… all eleven passengers plus the porter(making a good English jury) stabbed the manonce each in retribution for his kidnapping andmurder of a baby girl many years before.… Inthe case of the end-Permian mass extinction amore complicated explanation may be requiredby the evidence.… I believe that the extinctioncannot be traced to a single cause, but rather amultitude of events occurring together …

Erwin and many other geologistswere yet to be convinced. Perhaps theSiberian Traps had a hand in theextinction, but they were not the sole

cause. Or were they?

The killing traps

In 1992, Ian Campbell of the AustralianNational University and his colleagues16

presented a killing model for theSiberian Traps. They argued that theeruptions led to extinction by theinjection of huge amounts of sulphurdioxide and dust into the atmosphere.Analogously with the KT impact model,the dust blacked out the Sun and causeddarkness throughout the time of theeruptions. Global darkness leads toglobal freezing, and widespread

extinction of life on land and in the sea.Campbell and colleagues presented fivelines of evidence:

• The Siberian Traps represent thelargest volcanic event on land of the past600 million years.

• The Siberian Traps are associatedwith huge copper-nickel-sulphide orebodies, and thus the eruptions wereprobably rich in sulphur dioxide gas.

• The erupting magmas passed upthrough sediments containing theevaporitic mineral anhydrite, a richsource of sulphur that would have addedmore sulphur dioxide gas.

• The basalts are associated with agreat deal of tuff, material flung out ofthe volcanic vent into the air.

• The tuffs contain large amounts ofrock from the underlying sediments.

The tuff and rock fragments wereunusual for a flood basalt eruption, andCampbell and colleagues argued thatthese materials proved that the Siberianeruptions had been especially violent,and thus capable of injecting vastamounts of dust and sulphate gases intothe atmosphere.

Volcanoes produce two main gases,sulphur dioxide and carbon dioxide.Sulphur dioxide is a ‘greenhouse gas’,

and it initially causes warming.However, it soon reacts with water inthe atmosphere to produce sulphateaerosols that absorb the sun’s radiationby backscattering, and hence causecooling. It is well known that carbondioxide, another greenhouse gas, causeswarming of the atmosphere by trappingthe heat from the Sun’s rays near theEarth’s surface. Increases in the amountof carbon dioxide in the atmosphereincrease the heat-trapping effect, andtemperatures worldwide increase.Carbon dioxide emissions, amongothers, from industry and from cars aresaid to be producing measurableincreases in temperature today.

In a critique of the model proposed

by Campbell and colleagues, TonyHallam and Paul Wignall noted thatthere is essentially no geologicalevidence for dramatic global cooling,the key effect of the hypothesis. All theevidence actually points to globalwarming. Nor does the timing seem tomatch. Lasting for nearly one millionyears, the eruptions took much longerthan the extinctions. And finally, Hallamand Wignall suggest that the tuffs androck fragments may have beenmisidentified. Field evidence suggeststhat they are more likely the products oferosion of recently erupted lavas whichwere transported in rivers during aquiescent phase between eruptions.

Despite these seeming flaws, many

geologists have been impressed by thelink between the eruption of the SiberianTraps and the mass extinction. PerhapsCampbell and colleagues were broadlyright, but they picked the wrong model. Itis important to tease apart the sequenceof global events in much more detail.What climatic changes were going on inthe latest Permian and the earliestTriassic, and can these be explained bymassive eruptions?

Global warming

For a number of years, Steven Stanley ofJohns Hopkins University has arguedthat most, if not all, mass extinctions

were caused by global cooling.17 Hisevidence was fourfold: that mainlytropical species disappeared; that theextinctions were gradual; that there wasno limestone deposition after the event;and that evidence for ice could be foundin the sediments.

But none of these observations istrue for the end-Permian event.Temperate and polar speciesdisappeared at a high rate and theextinction was geologically rapid. Thereis extensive limestone formation in thetropical belt throughout the EarlyTriassic, showing that climates werewarm to hot. Also, there is no very goodevidence in the rocks for icy conditions,except near the then north and south

poles – which is not particularlysurprising.

Perhaps, suggest Tony Hallam andPaul Wignall, the end-Capitanian eventmight have been associated with globalcooling. This was the smaller massextinction that happened five to tenmillion years before the end of thePermian. Only tropical species wentextinct at that time, there was very littledeposition of limestone after the eventand there is sedimentary evidence ofglaciation. Indeed, the glacial tillites ofSiberia and Gondwana are well known.Tillites are fossilized tills, the masses offine debris and rocks commonly calledboulder clay, deposited beneath, or infront of, a glacier. There was also a

marine regression at the time which sawthe end of the Guadalupian reefs ofTexas. Perhaps this lowering of sealevel was related to glaciation: as iceaccumulates around the poles, water isfrozen into the ice and withdrawn fromthe sea.

Tony Hallam and Paul Wignall makea strong case18 for the exact opposite –global warming – at the end of thePermian. In their book, the appropriatesection is headed ‘Global warning’, aninteresting typographical error.Evidence for a dramatic increase inglobal temperatures comes from theoxygen isotopes, from the sediments andfrom the record of plant life.

As we saw in Chapter 7, oxygen

isotopes can be used as apalaeothermometer, and there was adrop in the value of the δ18O ratio ofabout six parts per thousand. Thiscorresponds to a global increase intemperature of around 6°C. This shift inthe oxygen isotope ratio was firstdetected in a borehole record across thePermo-Triassic boundary atGartnerkofel in Austria,19 and it hasbeen confirmed by measurements fromSpitsbergen and elsewhere. An increaseof 6°C may not sound much, but it wouldmake profound differences to thedistributions of plants and animals onland and in the sea. Recall thatclimatologists have been getting veryexcited recently about an increase of half

a degree in global temperatures duringthe past century.

The sediments seem to point toincreasing temperatures too. Palaeosols,literally ‘ancient soils’, can be a verygood thermometer of past climates. GregRetallack of the University of Oregon,who suggested that he had found shockedquartz at the Permo-Triassic boundary ofAntarctica and Australia, has shown amajor shift in soil types across theboundary.20 For example, Late Permiansoils in those regions are peats,suggesting cool conditions and winterfreezing, such as seen today in Siberiaand in northern Canada (latitudes 50–70°N). These are followed, however, byearliest Triassic warm temperate

palaeosols, more comparable to modernsoils seen in Ohio and Mississippi andin central Europe (latitudes 35–50°N).Rich, productive swamp soils of theLate Permian disappear entirely. Theshift from cool to warm conditions isconfirmed by an increase in theweathering of the rocks and increasedrainfall. Retallack argues that his studiesin Antarctica and Australia confirm thedevelopment of a ‘high-latitudegreenhouse’ in the earliest Triassicsouthern polar regions.

‘Absence of evidence is notevidence of absence’, as we alwaysteach our students. None the less, thereis no evidence of ice anywhere in theworld during the Early Triassic; no

tillites, no scratches or U-shaped valleysformed by putative glaciers, and nodropstones. Dropstones are often ratherdebatable. They are a genuinephenomenon, which occurs whenpebbles and boulders, stripped from theland by a glacier, are carried out to seaand fall to the sea floor as the base of theiceberg melts. The difficulty is inrecognizing them in the ancient rockrecord. One person’s dropstone isanother person’s odd or inexplicablepebble. Still, the absence of evidence ofice recorded in Early Triassic sedimentscan be taken as evidence that there wasno ice – until someone finds a counter-example of course.

Stronger evidence for global

warming comes from the record of fossilplants. The tree-like seedfernGlossopteris, widespread throughoutGondwana, died out at the end of thePermian, as did the other cold-adaptedplants associated with it. They werereplaced instead by warm temperatefloras dominated by the conifers Voltziaa n d Voltziopsis. In contrast to thediverse Glossopteris floras of the latestPermian, the new floras weredepauperate, meaning that there wererelatively few species. The fewsurviving plants were also designed tocope with hostile habitats, beinggenerally low-growing, and havingadaptations to withstand difficultgrowing conditions. Life was tough in

the ‘post-apocalyptic greenhouse’, asRetallack terms it.

Global superanoxia

There is a dramatic change also inmarine sediments at the Permo-Triassicboundary, as we saw. In many locations,ranging from the tropics to the poles, thediverse and fossil-rich limestones,mudstones and sandstones of the latestPermian switch to monotonous blackmudstones, containing very few fossils.This is a classic shift to anoxia. Sincethe shift is so large, and so pervasive, itcan justifiably be termed ‘superanoxia’.

The colour of mudstones is a goodguide to their conditions of deposition.Our students learn the mantra, ‘red isoxidized, green is reduced, black isanoxic’ (this is broadly, but notcompletely, true). The colours dependon the mineral content, and the mineralcontent can relate to conditions at thetime of deposition. The red/green colourpairing reflects the kind of iron oxide inthe mudstone. Red colours come fromhaematite, produced in the presence ofoxygen, just like rust. Green, or grey,colours indicate a different state of ironcalled goethite, which has lost oxygen,either at the time of deposition, or afterburial, by a process termed reduction.

Black mudstones mostly obtain their

colour from high proportions of carbonin organic matter. Organic matter canonly survive in the absence of oxygensince under normal conditions, withoxygen present, it would be consumedby microbes and animals. Even if thescavengers are not there, the carbonwould rapidly combine with oxygen toform carbon dioxide.

Black mudstones commonly followthe Permo-Triassic boundary. This wasseen in the succession at Meishan inChina, and others nearby, where theearliest Triassic black shales alsocontain pyrite, more commonly known asfool’s gold. Pyrite, or iron sulphide,forms only in anoxic conditions, such asin the stinking black slime at the bottom

of stagnant pools, but more commonly insimilar masses of rotten organic matteron the sea floor.

Anoxic mudstones are not apeculiarity of China alone. Paul Wignalland Richard Twitchett21 of theUniversity of Leeds have examinedPermo-Triassic boundary sectionsthroughout Asia, Europe and in northernlatitudes, and they find the samephenomenon. Spitsbergen, for example,lies to the north of Norway and to theeast of Greenland, at 78°N, and it wasalso in north polar latitudes in thePermo-Triassic. Wignall and Twitchettfound abundant evidence for earliestTriassic anoxia there: a switch to blackshales, the presence of pyrite, the

absence of burrows and the scarcity ofbottom-living organisms.

The association of fossils andsediments can be revealing. Undernormal conditions in shallow seas,numerous organisms live in and on thesea bed. Corals and other reef-buildersare fixed to the sea floor. Gastropodsand arthropods creep about looking forfood. Worms, bivalves and otheranimals make burrows into the sea-bedmuds and sands, either for protection orin search of food. The latest Permiansediments show evidence of all of thisbiotic activity. Then it stops. Theearliest Triassic black mudstonescontain no corals or other fixed sea-floor beasts. Burrowing is absent. The

only living things are some fishes thatswam high above the sea floor, and thepaper pectens such as Claraia. As wesaw in Chapter 8, Claraia, andassociated forms, seem to have beenspecialized in living in low-oxygenconditions.

Global superanoxia was associatedwith a crash in productivity in theearliest Triassic. Carbon isotope curvesshow a dramatic negative swing exactlyat the Permo-Triassic boundary, asshown in the Meishan section in China(Chapter 7), and in every other Permo-Triassic succession that has beenstudied, whether marine or continental.The isotope curve swings back tonormal values within a few hundreds of

thousands of years after the massextinction event. What was going on?

All kinds of explanations have beenproposed for the carbon isotope shift.The negative shift, reflecting an increasein release of carbon-12, could have beencaused by erosion of Permian coals.These coals, formed from Glossopterisand other forest plants, locked up largeamounts of organic carbon-12. Aftererosion, the carbon-12 might have beenwashed into the sea, and accumulatedfast enough to cause the carbon ratio toshift. But the dramatic reduction in thecarbon isotope curve right at the Permo-Triassic boundary could indicate acollapse in productivity.

Productivity is a measure of

biological activity, indicating therichness of life and its ability to processcarbon and other materials. Aproductivity collapse occurs when mostliving things die. The surface of the landwould be covered in the trunks of deadtrees, and mounds of stinking, rottenplants and animal carcasses. The seafloor would be awash with dead corals,shellfish and fishes. The carbon-12locked up in these plants and animals issuddenly incorporated into the sediment,and the ratio of carbon-12 to carbon-13,the normal inorganic state, shiftsdramatically in the direction of carbon-12. But would the death of all life at thetime be enough to explain the negativeshift of four to six parts per thousand in

the δ13C ratio?So how does this all work? Eruption

of the Siberian Traps led to globalwarming, superanoxia and a productivitycrash? Or was there more?

The methane burp

It all depends on a rise in carbondioxide levels. What happens in theatmosphere happens in the oceans aswell, as changes in the gas compositionor temperature of the air penetrate theupper levels of the ocean. Also, changesin productivity and the cycling oforganic carbon on land also affect the

situation in the sea, as material is fed invia rivers.

Single-cause models for catastropheare persuasive. And yet traditionallytrained geologists are a cautious breedsometimes. In 1997, in their book onmass extinctions, Tony Hallam and PaulWignall22 present detailed accounts ofthe Siberian Traps and of the majorclimatic changes associated with theend-Permian event. Because of problemsof matching the timing of events, andbecause they were uncertain whetherthese fissure basalt eruptions could havebeen explosive enough to inject thenecessary dust and gases high into theatmosphere, they conclude that theeruptions did not trigger the extinctions.

In their flow chart explaining thesequence of events, the Siberianeruptions do not figure. Four years later,Paul Wignall has included the eruptionsas the major cause, but with somecaveats. Why the shift of opinion?

The main concern in 1997 was thatthe eruptions themselves could not havesupplied sufficient carbon dioxide intothe atmosphere to cause the globalwarming of 6°C and the levels ofsuperanoxia that had been detected.Other sources were needed. In 1997, allthe postulated carbon dioxide came fromoxidation of coals in southernGondwana. If huge amounts of coal areexposed to the air, much of the carbon inthe coal will combine with oxygen in the

air to produce carbon dioxide. But thissource has always been problematic.Could it really produce enough carbon?And at a fast enough rate?

It’s not enough simply to find a newsource of carbon dioxide; that source hasto be capable of overwhelming the usualatmospheric feedback systems. Undernormal conditions, carbon dioxide isproduced by animals breathing out, byburning of wood and fossil fuels, andfrom volcanoes. Negative feedbacksystems have a tendency to cope withfluctuations and to bring excesses of oneinput back to a standard level. ‘Negativefeedback’ means that a process iscountered by the opposite, or negative,process, so regulating the effects of the

process and maintaining a steady state.‘Positive feedback’, on the other hand,means that the process is enhanced bymore of the same, with further positiveprocesses operating in the samedirection.

So with the atmosphere. Excesscarbon dioxide is mopped up by plants(during photosynthesis plants absorbcarbon dioxide and produce oxygen) andthrough weathering. Carbon dioxide isstripped out of the atmosphere by rainwater, forming weak carbonic acid,which then dissolves limestones on theground. The carbon combines with theweathering products of the limestone,and the oxygen is given off as carbondioxide. If you drip an acid on to

limestone, the limestone will fizz – thisis the carbon dioxide bubbling off.

By 2001, a trendy new carbonsource had been identified. And this wasone that was fast. Gas hydrates arecrystalline solids composed of a cage ofwater molecules trapping gas inside.The water cages can trap various gasmolecules, including carbon dioxide andhydrogen sulphide, but the commonestgas hydrates trap methane, a gascomposed of carbon and hydrogen. Gashydrates form at water depths greaterthan 500 metres, and particularly inpolar regions. Because of the highpressures at such depths, the gashydrates are amazing gas concentrators;if 1 million litres of methane hydrate is

brought to the surface suddenly, 160million litres of gas can be released.

Since the 1970s, when gas hydrateswere discovered, they have beenidentified deep in sediments around themargins of most continents, andparticularly around the poles. The hugefrozen masses of ice and compressed gasfill pore spaces within the sea-floorsediments, and occupy vast fields thatcan be detected by means of geophysicalsoundings. Worldwide, it is estimatedthat gas hydrates contain at least 10,000billion tonnes of carbon, about twice theamount of carbon held in all fossil fuelson earth. If some perturbation hits one ofthese gas hydrate bodies, and the gas isreleased, huge volumes of carbon

dioxide or methane would bubble upthrough the ocean and explode from thesurface, temporarily displacing thenormal atmosphere above.

Could this be a killer? On 3December 1872, the ship Marie Celestewas found adrift off the Azores. Therewas no sign of life on board, eitherabove or below decks. There were noclues to explain why the crew haddisappeared. Indeed everythingappeared to be quite normal. In thecrew’s quarters, clothing lay foldedneatly on bunks and washing hung onlines; in the galley, breakfast had beenprepared and some of it had beenserved. Could the crew have been killedby the release of a massive bubble of

methane hydrate? Millions of cubicmetres of methane or carbon dioxideerupting from below would have adevastating effect, but would then beswallowed up into the atmosphere,leaving no trace. Why were all the crewmissing? This will never be known –perhaps the pulse from below and thestagnation of the atmosphere made themall run on deck and jump overboard insearch of fresh air.

What if numerous gas hydratebodies, all round the world, were tohave been released at the same time?Evidence has now been found for such amass gas escape, a so-called methaneburp, 55 million years ago, at the end ofthe Palaeocene. At that time there was a

pulse of global warming, by 5 to 7°Cover approximately 10,000 years, asshown by oxygen isotopes and therecord of fossil plants. It has beensuggested that this pulse of warming wascaused by the release of 2000 billiontonnes of methane hydrate into theatmosphere.

The pulse of warming was brief, andconditions rapidly returned to normal.Gerry Dickens of the University ofMichigan and colleagues suggest23 thatthis is good evidence that the warmingwas caused by gas hydrates. The rapidheating led to the death of many species,the excess organic matter from deadplants and animals was washed into thesea, and carbon dioxide levels in the

atmosphere were quickly reduced by theincorporation of the organic matter intooceanic deposits and by increasedweathering following the loss of plantcover.

The end-Palaeocene methane burpdid not lead to a major extinction event.The effects were worldwide, and manyspecies died out, but the Earth returnedto normal soon enough, and most speciesrecovered. Could such a model beenough to kill almost all life?

Explaining the carbonisotope spike

The gas hydrates were probably not themain killer at the end of the Permian. Butthey may help to explain the massivenegative shift in the carbon isotopecurve, which dropped by four or fiveparts per thousand. This perhaps doesnot sound much of a shift towards thelighter isotope of carbon, carbon-12; butit actually represents a global shift in theentire carbon budget – the introductionof billions of tonnes of light carbon intothe oceans.

What are the possibilities? Theinflux of isotopically light carbon couldhave come from the collapse ofproductivity that happened at the Permo-Triassic boundary, and the entry of hugeamounts of rotting wood and animal

carcasses into the sea. But, Paul Wignallhas calculated, this would not be enough.It is estimated that the entire biomass oflife on Earth today contains 830 billiontonnes of carbon. If all life is killedinstantaneously, that amount of organiccarbon-12 could be washed into the seaand buried. But there are already some50,000 billion tonnes of inorganic,heavier carbon-13 in the ocean-atmosphere system, so the addition of830 billion tonnes, less than a 50th of theamount, would make very littledifference to the ratio.

Even the Siberian Traps eruptionscould not have supplied enoughisotopically light carbon. If the volumeof basalt produced was 2 million cubic

kilometres, that would have produced10,000 billion tonnes of carbon, whichwould have been a mixture of carbon-12and carbon-13. So, again, it was notenough to cause the carbon isotope shift.Even with the best figures, the SiberianTraps eruptions could have producedonly 20% of the shift that actuallyhappened.

Geologists have embraced gashydrates with fervour – almost with asigh of relief. When the calculations aredone, nothing else has enough lightcarbon, nor can act fast enough. Thecarbon in gas hydrates is isotopicallyvery light, with a δ13C value of -65 partsper thousand. The release of only 10%of the estimated 10,000 billion tonnes of

carbon contained in gas hydrates todaywould be sufficient to cause the shift inthe δ13C ratio by -5 parts per thousand.The secret is the very light compositionof carbon. Although the Siberian Trapsmay have released the same mass ofcarbon, its isotopic weight was muchheavier, and could not have producedthe observed negative spike.

The killing model

Paul Wignall has put everything togetherinto a single flow chart (Fig. 41). Thekey crisis seems to have been theeruption of the Siberian Traps.

Worldwide devastation was caused bythe production of different gases duringthe eruptions, and these gases werepresumably pumped into the atmospheresporadically during the entire span of theeruptions. Perhaps a single majoreruption could have been absorbed bythe Earth, and the short-term disturbanceof the atmosphere-ocean systemcorrected by normal feedback processes.But repeated eruptions may have beentoo much, and may have led inexorablyto total collapse of all normalinteractions between the physical worldand life.

41 The cycle of death. A summarydiagram showing how the eruptionof the Siberian Traps led to majoratmospheric changes and to thecollapse of most of life on Earth251 million years ago.

Four gases from the eruptions mayhave been to blame. Carbon dioxide hadthe longest-term effects, leading

immediately to global warming andanoxia, which persisted for hundreds ofthousands of years. Each pulse oferuption may have reprimed the effects,and prevented any normal feedbacksystems from kicking into operation. Therelease of gas hydrates added to themisery.

Sulphur dioxide was also produced.This gas has a shorter residence time inthe atmosphere, but the cooling effects ofthe sulphates may have caused a snapglaciation in some parts of the world,with associated falls in sea level asmarine water was frozen into ice.Whether there was such a glaciation, andhow long it lasted, cannot be said atpresent. And, as we saw above, there is

limited geological evidence for freezing,but if it was short-term, as the theorysuggests, one would not necessarilyexpect to find evidence preserved in therocks.

Chlorine gas may also have beenproduced. In conjunction with thesulphates and the carbon dioxide thesewould produce acid rain. Whencombined with water, these gases formhydrochloric acid, sulphuric acid andcarbonic acid, and hydrofluoric acidmay also have been released. If such adelightful cocktail of acids were to rainout of the sky, normal plant life wouldhave been devastated, just as today acidrain kills forests. With normal plantsdramatically reduced, animal life on

land would go too. Perhaps thispostulated acid burst wiped out much oflife on land and led to the fungal spike,the mushrooms and moulds being thefirst land life to be able to recover.

Acid rain also, of course, increasesthe rate of normal weathering on land,and the loss of plants would make itworse as soils were stripped off.Retallack and colleagues certainlydetected this in the record of soilsacross the Permo-Triassic boundary, andincreased rates of runoff of sediment intothe sea are indicated also by a shift instrontium ratios. A dramatic increase inthe ratio of strontium-87 to strontium-86across the Permo-Triassic boundarysuggests that huge amounts of terrestrial

material were entering the sea viarivers.

The whole end-Permian crisis mayhave been made even worse by arunaway greenhouse effect. Normally,the atmosphere-ocean system willcorrect imbalances, and return carbonand oxygen levels to normal. This is anegative feedback process. If carbondioxide levels increase, burial oforganic matter, weathering orproliferation of forests will eat up theexcess gas. However, a runawaygreenhouse is a positive feedbacksystem. An increase in carbon dioxide,for example, is not countered byprocesses that mitigate the effect. On thecontrary, it triggers processes that add

yet more carbon dioxide to theatmosphere.

The end-Permian runawaygreenhouse may have been simple.Release of carbon dioxide from theeruption of the Siberian Traps led to arise in global temperatures of 6°C or so.Cool polar regions became warm andfrozen tundra became unfrozen. Themelting might have penetrated to thefrozen gas hydrate reservoirs locatedaround the polar oceans, and massivevolumes of methane may have burst tothe surface of the oceans in hugebubbles. This further input of carbon intothe atmosphere caused more warming,which could have melted further gashydrate reservoirs. So the process went

on, running faster and faster. The naturalsystems that normally reduce carbondioxide levels could not operate, andeventually the system spiralled out ofcontrol, with the biggest crash in thehistory of life.

The view from the burrow

What did all of this look like at the time?Imagine the scene in the Karoo Basin inDicynodon Zone times, which weencountered in Chapter 9. Dicynodonitself (Fig. 42), the medium-sized plant-eater that was most abundant at the time,may have been able to make burrows inwhich it could escape from the normal

rigours of the tropical-monsoonalclimate in which it lived. As the crisisapproached, Dicynodon would havescuttled along the river bank and plungedinto his cool burrow, expecting that itwould all pass in a day or so and hecould crawl out again.

The first basalt eruptions beganthousands of years before, and far away,in Siberia, and continue, sporadically.None of the noise of the explosionswould be heard in Africa, nor wouldDicynodon have seen any of the eruptinglava, ash or gas. But air temperaturesmight bounce up a little. Locally, aroundthe eruption site, there might be a snapfreeze caused by the emission of sulphurdioxide, but that would be a short-term

phenomenon, soon overwhelmed by thewarming effects of the carbon dioxideemission. The first eruptions pass prettywell unnoticed.

Then, a year or so later, there is alarger eruption. A further snap freeze isreplaced by a greater rise in airtemperature. This time Dicynodon feelsit. It is the dry season, the time betweenthe annual monsoonal rains, and it hurts.Life is balanced on a fine marginbetween survival and death during thedry season in any case, as we saw inChapter 9. Even a one degree rise intemperature can kill off more plants thannormal, and then more herbivorousanimals fail to survive through to therainy season. The blast of heat sends

Dicynodon into his hole.This time, the eruption initiates some

further processes. The cocktail of gasesejected into the atmosphere rises highinto the stratosphere and encircles theglobe. The gases and fine dust distort thenormal appearance of the heavens –sunrises and sunsets look weird, withsplashes of red, yellow and purple, andthis is seen all round the world. A fewdays later, the perturbation triggerscatastrophic acid rain. The chlorine,fluorine, sulphur dioxide and carbondioxide emitted from the volcanocombine with rain water in the highclouds to produce a cocktail ofhydrochloric, hydrofluoric, sulphuricand carbonic acids. For millions of

square kilometres around the eruptionsite, the acid rain burns off most of theplants. First to go are the trees andlarger plants. Even as far away as SouthAfrica, the effects can be seen. Plants liedead where once they grew. They rotand decompose. Dicynodon creepsabout, looking for some palatablemorsels, but finds very little – just somemushrooms, mosses, ferns and club-mosses nestling in damp crevices aroundthe river banks.

42 Dicynodon explores thedevastated post-apocalypticlandscape.

Then comes a distant rumbling,unheard in South Africa, but the coup degrâce none the less. Since the eruptionsbegan, some 10,000 years earlier, theatmosphere and sea surface havewarmed by two or three degrees, and thefrozen northern polar region begins tomelt around the fringes. The polarregions in the Permian were muchsmaller than they are today, with limitedice caps. But frozen tundra extendshundreds of kilometres away from thepoles, and the ocean margins are frozentoo. A great icy mass of gas hydrate,

locked in the sediments at the margin ofthe polar sea, is warmed by a degree ortwo, and it suddenly gives way. First afew bubbles, then many, and finally ahuge expansion of gas – 160 times theoriginal volume. What was once frozenand at high pressure, becomes gas atnormal temperatures and pressures, anda vast volume of methane and carbondioxide bursts upwards through theoceans and shoots out into theatmosphere, raising spouts of seawaterhundreds of metres into the sky.

There had been a long-term declinein the proportions of oxygen in theatmosphere through much of the Permian,and it reached a low point at the time ofthe extinction. In the course of a few

days, Dicynodon, feebly searching forscraps among the stinking decay ofplants in southern Africa, and alreadyfeeling a rise in temperature of two orthree degrees, is now hit by a furtherdevastating blow. Normal levels ofoxygen are being driven downwards –he is gasping for air. And, day-by-day,the asphyxiation becomes worse.

After a week, heavier rains come.The monsoon has begun. The rain is stillacidic, although much of the acid hasnow been washed out of the system. Andthe rain carries away all the stinkingvegetation down the slopes, into therapidly filling wadis. Jostling treetrunks, branches and mats of leaves rushdown to the sea, where they are dumped

a few kilometres offshore, at the end ofthe estuarine tracts. But most of the soilis washed away too. Without the bindingroots of the plants, the soil is vulnerable.After a few days, there is almost noearth left, just bare rock, with pockets ofsoil clinging on in hollows wheremosses, ferns and club-mosses havesurvived. The countless billions oftonnes of organic carbon locked into theplants and the soil across the wholeKaroo have been stripped into the sea.Almost nothing remains. With the soilwent the worms, spiders, centipedes,flies, beetles and everything that couldnot hang on to the rocks and the rushingtorrents of water.

Carbon dioxide levels in the

atmosphere are still higher than normal.And there are no negative feedbackprocesses. Normally, the carbon dioxidewould be removed from the atmosphereby photosynthesis, and the monsoonalrains would have been followed by adramatic greening of the land. Dried-uptrees would spring into life, producingleaves from their gnarled branches. Thebare earth would miraculously sproutlow ferns and seedferns, as dormantseeds broke into life. But the soil hasgone, the land is just naked rock.Nothing like this had happened since thePrecambrian, some 300 million yearsbefore, when life on land had not yetevolved.

Dicynodon follows the water

downhill to the sea, half-starved now,having gone without food for more than aweek. Mushrooms, which seem to be allthat can flourish, are the only thingavailable, and they are far from hispreferred diet. His instincts suggest therewill be food where the water is. But heis wrong. Everything is topsy-turvy inthis apocalyptic world. Just as the plantson land were killed by the acid rain, sotoo the seaweeds that fringed the shores.The increased carbon dioxide levels inthe atmosphere have penetrated the topdozens of metres of the sea, and theplankton has been decimated. Within aweek, all the shoals of fishes that used torely on the plankton as their staple diethave starved too; then on up through the

food chain: the sharks and larger fishesthat fed on the planktivorous smallerfishes die too. The whole sea ispoisoned. Rising temperatures haveimposed anoxia.

Dicynodon, and all the other animals– the closely related smaller and largerdicynodonts, the bulky, knobbly plant-eating pareiasaurs, the small scuttlingprocolophonids, therocephalians,millerettids, and the large, sabre-toothedgorgonopsians – are close to death. Theywander about on a blank, rockylandscape. They have difficulty inbreathing, and their backs are burnt atmidday by the hotter-than-normal sun.

Then comes the third eruption. It isnot by itself a particularly huge eruption.

But it leads to a further cycle of acidrain. Temperatures rise again by afurther fraction of a degree. Another vastbubble of methane and carbon dioxide isreleased in the far north from a frozengas hydrate deposit. Dicynodon is nowliving through a runaway greenhouseeffect. Nothing can turn back thedevastating rise both in carbon dioxideand in temperature. He curls up and dies,along with nearly every other livingthing on land.

In the sea, the vast influx of organicmatter from the land – all the deadplants, animal carcasses and soil – havecarpeted the seabed in a stinking, blackslime. The decaying organic matterconsumes oxygen and gives off hydrogen

sulphide. Seafloor life – all the reefs andtheir denizens, as well as the creepingworms and arthropods, and theburrowing molluscs and shrimps – die.Their carcasses are incorporated into thefetid, black, anoxic bottom slime.Stripped of oxygen from the atmosphereabove, and with the black mud below,the oceans go into a spiral of anoxia:oxygen levels fall, step-by-step, untilalmost nothing survives. Only a fewworms, brachiopods and molluscs thatcan exist at depth in oxygen-poorconditions manage to live through theseharsh conditions.

The eruptions continue, at randomintervals, to pulse basalt lavas overSiberia. Hundreds of metres of fresh

rock accumulate. Sometimes eruptionsare separated by days or weeks; at othertimes, there may be a standstill for a fewthousand years, and the rare survivingplants and animals manage to re-establish themselves for a short time,before a further cycle of devastationbegins. Some of the eruptions are smalland have limited global effects, ofcourse, but others are large enough tolead to the global effects just described.Some day, with ever-more precise study,it may be possible to tease apart some ofthe detail of the separate phases of theeruption of the Siberian Traps, andwhether the killing happened all at thestart of the eruption cycle, or whether theprocess was drawn out over half a

million years.

Is this what reallyhappened?

All the pieces of the cataclysm 251million years ago have been put in place.Research in the past ten years has led toan astonishing, earthbound scenario foralmost complete devastation of the Earthand of its inhabitants. The killing modelmakes sense in terms of what is nowunderstood about eruptions and theireffects on the atmosphere, about oxygenand carbon cycling in the earth-lifesystem, about the composition of the

oceans, and about gas hydrates. Butmuch of this is very new work, and itmight have to be modified in the future.

For many, there is a lingering desirefor something more apocalyptic, moreinstantaneous, some deus ex machinasuch as a huge extraterrestrial impact.Surely, they argue, if the KT event, when50% of species disappeared, required avast meteorite, we need something evenmore catastrophic to kill off 90% ormore of species? But the evidence forimpact at the Permo-Triassic boundaryis limited at the moment. That might allchange of course any day, if thehelium/fullerene story is confirmed, if acrater of suitable age is discovered, or iflarge amounts of shocked quartz and

iridium turn up in rock successions indifferent parts of the world. However,I’ll bet on the Siberian Traps coupledwith gas hydrates for the moment.

What was so special about theSiberian Traps, and the whole end-Permian scene, that could have allowedthe crisis to develop? The SiberianTraps were not the biggest flood basaltsof all time. Much larger flood basaltvolumes were erupted over the CentralAtlantic, Java, the Caribbean-Colombianarea and the Brito-Arctic province, 200,120, 90 and 60 million years agorespectively. But these four wereassociated with generic extinction ratesof only 5–30%, notably at the higher endof the range, but hardly devastating, and

certainly not on a scale with the end-Permian loss of genera.

Maybe it was simply a coincidenceof factors. The end-Permian was theonly time when the continents were fullyassembled into a single supercontinentand when there was a large flood basalteruption episode. During the later large-scale basalt eruptions, the continents haddrifted apart and maybe life haddiversified sufficiently in the differentcontinents and oceans to be able to resista range of severe climatic changes. Onealmost certainly has to add thecoincidence of a massive methane burp,although there is no independentevidence yet for this (apart from thedifficulty of otherwise explaining the

remarkably large and rapid negativecarbon isotope shift). Thorough testingof all the hypotheses by Robert Bernerof Yale University has shown24 thatmassive methane release has to be themain cause of the dramatic carbon shift,with lesser contributions from volcanicdegassing and mass mortality. This isn’tproof, but he ran all the possible causesagainst his well-established climaticmodels, and very few doubt Berner’spre-eminence in this arcane world oflarge-scale number-crunching.

Much more has yet to be found outabout the end-Permian crisis. Geologistsand palaeontologists are far fromunderstanding step-by-step just whathappened. But, as we have seen, ideas

have sharpened and focused remarkablysince 1995, and will doubtless continueto do so. One thing is clear, however.The biggest mass extinction of all timedid happen 251 million years ago, andeven if we cannot yet fully explain why,it is important to look at theconsequences of cutting life down to10% or less of its normal diversity.There are lessons to be learnt.

12

THE SIXTH MASSEXTINCTION?

In 1992, Al Gore, then the Vice-President of the United States ofAmerica, wrote,

… it occurred to me that … we are causing 100extinctions each day – and many scientistsbelieve we are …1

This is a startling figure, and theprediction resulting from thecalculations quoted by Gore is that all oflife will be extinct in 400 to 800 years.Do we believe this? What is thescientific basis for such dramaticpredictions? Or should we settle backcomfortably with the extreme Bible-beltAmericans who think that everything onEarth was created by God for the benefitof humanity, and therefore that anythingdone by human beings is by definitiongood?

Probably both positions are grosscaricatures, and it would be sensible tobe cautious. Al Gore based his statementon the reputable calculations of Paul andAnne Ehrlich in 1990 that perhaps 70–

150 species are becoming extinct eachday.2 Scaling this up to current estimatesof total global diversity leads to theabove alarming prediction of how longlife will last on the Earth. Perhaps thedaily extinction estimate is too high (seebelow), but such calculations alwayslead to startling conclusions about thetotal time to extinction of all life.

If human activities are truly causingsuch devastation, then we are certainlywitnessing the sixth mass extinction (theother five being the geologicallydocumented ‘big five’). In this case, aclose understanding of the events of thepast will shed light on what is happeningnow, and what may happen in the future.In particular, people often ask, how long

does it take for life to recover after thedevastation of a mass extinction?Recovery after the end-Permian event isdocumented, and palaeontologists cangive some firm answers to that question.

But what about Al Gore’s alarmingsuggestions? Are such estimatesreasonable?

Modern diversity: 2 millionspecies?

Current estimates of biodiversity, thenumber of species on Earth, range from2 million to 100 million. It seemsincredible that we frankly have so little

idea about how diverse life is today.The low-end figure of 2 million,

quoted still – though rarely – by someconservatives, comes from the mostsolid evidence of all, a count of everynamed species. By the year 2000, some1.8 million species had been named inthe scientific literature, the majority ofthem, about 1 million, as it happens,being insects. It is hard to give an exacttotal, since new species of plants andanimals are described in any of thethousands of scientific journals of theworld, including publications ofmuseums and institutions small andlarge. There is, as yet, no centralrepository of such information, althoughthere are plans afoot to marshall all the

data into a central electronic storagesystem on the World Wide Web,available to all.

Do we stick to 1.8 million as ourestimate of current biodiversity? Yesand no. The number is obviously far toolow for one reason, and rather too highfor another. Too high because ofsynonymy. Synonyms in biology, as incommon speech, are different terms forthe same entity. Despite all their bestefforts, biologists are in fact only humanand they make mistakes. Theirenthusiasm for finding new speciesmeans that they may mistakenly invent aname for something that has already beennamed by someone else. This mayhappen even when they know of all

previous work in their field. It may bebecause an earlier name was given insome particularly obscure journal, or theoriginal description might have beenincomplete, poorly illustrated ormisleading. Be that as it may, this leadsto a constant rate of synonymy, therenaming of species that have alreadybeen named. Fortunately, scientists arecritical folk, and they delight inidentifying the errors of others. Sooneror later, synonyms are rooted out andexposed. But that is a part of science: allpublished work is open to scrutiny andchecking.

More important in many ways thanmerely naming new species,systematists, the biologists who work on

the relationships and diversity of life,constantly revise and restudy wholegroups. They might decide to take inhand a family of orchids or waterbeetles. In doing so, they travel theworld, look at type specimens (thespecimens that were selected asrepresentative of a new species when itwas named, sometimes called the name-bearer, and preserved in a museum forlater study), and produce comprehensiveschemes of characters and relationships.Then they may identify synonyms, andformally subsume such synonyms underthe names originally given. The globalrate of synonymy is about 20%, and italways has been. In other words, aboutone-fifth of new species announced to

the world by their proud parent(s) willinevitably be deleted on later revision.Our figure of 1.8 million named livingspecies then falls to about 1.4 million.

This, though, is almost certainly agross underestimate. New species arebeing found all the time, and even thoughone-fifth of them may be spurious, four-fifths are not. How many new species? Ifyou read the serious newspapers, newspecies of birds or mammals makeheadlines. A new mammal or bird isdiscovered only quite rarely, maybe oneor two a year, and usually after diligentexploration in remote parts of the world.This might suggest that we really knowpretty well all birds and mammals alivetoday, some 7000 and 5000 species

respectively. The story is different forinsects though: entomologists come upwith a regular 7000–8000 new specieseach year, and the only limitation on thisrate of discovery seems to be the numberof entomologists available and their rateof work. Entomologists that I know feelthat their situation is desperate: if theycould only work another hour per day,they could identify another dozen newspecies a year. Human frailty holds themback, and they lose sleep over it.They’ve named a million species so far– but there is no sign of an end to theirnecessary labours.

These contrasts can be appreciatedfrom a collector curve. The collectorcurve is a well-established way of

answering the question: ‘When should Istop collecting?’ It is simply a graph ofthe number of new species discoveredplotted against effort. In the classic case,a biologist is sent to some remote placeand is asked to compile an inventory ofspecies. How long should she take? Ifshe collects for a day, every plant or bugshe picks up will be something new toher. After a week, she knows all thecommon species, and only rarely finds anew form. The law of diminishingreturns sets in. Plotted as a graph (Fig.43), this shows that the rate of discoveryis high at first, and then rapidly tails off.The right-hand part of the line is tendingtowards a final figure, which it maynever reach (this is technically called an

asymptote), but it is clear roughly whatthat total is. If this collector curve thenrepresents the world-wide efforts ofsystematists in identifying new species,mammalogists and ornithologists arewell up on the asymptote, and they canpredict with confidence that they knowpretty well all there is to know.Entomologists, on the other hand, arestill struggling up the steep part of thediscovery curve, and it is impossible topredict when the rate of discovery willslow down. A rough estimate from thecollector curve of all organisms mightsuggest that the total number of speciesalready described and to-be-describedwill reach 5 million.

Insects are all very well. We see

them around us, and they are easy tocollect. But what about microbes – allthe bacteria, viruses, algae, protozoansand other micro-organisms that we can’tsee? Microbiologists just smile whenentomologists announce their problems.They haven’t even begun to scratch thesurface of current microbial diversity.Similarly amused are the systematistswho specialize in deep-sea organisms.They can only study the greyish mangledremains hauled up from the depths intrawls and sediment punches. Their taskis like that of an extraterrestrial whoswoops high over the Earth, and drops asample grab down from time to time, andhauls back into his spaceship whateverhas been encountered.

What about the meiofauna too, I hearyou ask? Indeed. The nearlymicroscopic, though not quite,transparent, soft-bodied bugs and worm-like creatures that live between sandgrains and among the soil particles, havebarely been studied. No one really knewthey even existed until the 1960s, norreally cared, and only a small number ofsystematists study them today. Perhaps aglobal inventory of named species is nota valid estimate of total globalbiodiversity. Another approach might beto adopt a sampling strategy.

43 A collector curve for thenaming of species. Systematists ofbirds and mammals have just aboutfound all the species that exist. Butentomological systematists aretrailing badly. This is not theirfault, since insects are thousandsof time more diverse than birdsand mammals. At full steam,entomologists are describing 8000or so new species of insects eachyear, and yet they seem to be nonearer finishing than they were100 years ago.

Bug guns: 100 millionspecies?

A sampling strategy is just what TerryErwin applied in the 1970s. In two well-known papers3 he outlined his method.He chose to look at tropical rainforestbeetles, to estimate their biodiversity ina small part of the world, and then toextrapolate from that to the whole world.He sampled the entire arthropod fauna –all the insects, centipedes, millipedes,spiders and their relatives – from thecanopy of the tree Luehea seemanniifrom Central and South America. Thiswas done by setting ‘bug bombs’ underthe selected trees, devices that pumppowerful insecticide upwards in clouds.When these had gone off, Erwincollected all the dead arthropods whichfell to the ground, and he sorted them

into species, many of them, as ithappened, never seen before.

Erwin estimated that there are 163species of beetles living exclusively inthe canopies of Luehea seemannii.There are about 50,000 tropical treespecies around the world, and if thenumber of endemic beetle species inLuehea seemannii is typical, thisimplies a total of 8.15 million canopy-dwelling tropical beetle species in all.This figure excludes forms that live inseveral tree species. Beetles typicallyrepresent about 40% of all arthropodspecies, and this leads to an estimate ofabout 20 million tropical canopy-livingarthropod species. In tropical areas,there are typically twice as many

arthropods in the canopy as on theground, giving an estimate of 30 millionspecies of tropical arthropodsworldwide.

Thirty million species of tropicalarthropods? If this were true then thatwould imply a total diversity of lifesomewhere between 50 and 100 millionspecies. Some wild-eyed biologistseven talked of figures of more than 100million! New work, published in 2002,has, however, challenged some ofErwin’s assumptions.

Mature reflection4 has suggested thatErwin’s estimate of total globalbiodiversity was still closer to the truththan more conservative estimates of 2–5million. Similar extrapolation exercises

have been performed for deep-seaorganisms, microbes, fungi andparasites, and they all point to totalglobal biodiversities of 20–100 million.Such estimates are astounding, and theyhave profound consequences. Shouldsystematists give up the endeavour todescribe and name all species since theywill never finish the task? Shouldgovernments employ many moresystematists in order to do the workproperly? How can conservationists andplanners begin to estimate the effects ofpollution and other human activities onbiodiversity, since no one has thefaintest idea how many species existtoday, nor what they are, and where theyare?

Modern extinction rates: 70species extinct per day?

The present rate of extinction can becalculated for some groups fromhistorical records. For birds andmammals, groups that have always beenheavily studied, the exact date ofextinction of many species is knownfrom such historical records – and,shamefully, not so historical in manycases. The last dodo was seen onMauritius in 1681. By 1693, it was gone,prey to passing sailors who valued itsflesh, despite the fact that it was ‘hardand greasie’. The last two great aukswere collected in the North Atlantic in

1844 – ironically, this pair wasbludgeoned to death on Eldey Island offIceland by natural history collectors.Some sightings were reported in 1852,but these were not confirmed.

Human activity has not simplycaused the extinction of rare or isolatedbirds. The last Passenger pigeon, namedMartha, died at Cincinnati Zoo in 1914.Only one hundred years earlier, the greatornithologist John James Audubon,reported a flock of Passenger pigeons inKentucky that took three days to go by.He estimated that the birds passed him atthe rate of 1000 million in three hours.The sky was black with them in alldirections. They were wiped out by aprogramme of systematic shooting,

which, at its height, covered thelandscape with their carcasses as far asthe eye could see.

These datable extinctions can beplotted (Fig. 44) to show the rates ofextinction of birds, mammals and someother groups in historical time.5 Thecurrent rate of extinction of birds is 1.75per year (about 1% of extant birds lostsince 1600). If this rate of loss isextrapolated to all 20–100 million livingspecies, then the current rate ofextinction is 5000–25,000 per year, or13.7–68.5 per day. With 20–100 millionspecies on Earth, that means that all oflife, including presumably Homosapiens, will be extinct in 800–20,000years. This is lower than the estimate

quoted by Al Gore, since it is based onrevised figures for the current dailyextinction rate. But the figures arestartling enough. This is the evidencethat we are living through the sixth massextinction. Is this reasonable? Or is it awild exaggeration?

44 The rate of historicalextinctions of species for whichinformation exists. Rates of lossmount in the twentieth century. Theapparent drop from 1950 to 2000is not an indication of animprovement: the figures wererecorded in 1990, and the full tallyof species lost up to 2000 has notyet been calculated.

Even if all of life does not go extinct,and it seems likely that some specieswould be tenacious, or lucky, enough, toescape whatever depredations andenvironmental catastrophes may bemeted out to them by human activities,can we foresee a recovery phase? If so,how does life recover after a mass

extinction?

Bouncing back on land

During the end-Permian crisis, life wasdriven down to 5 or 10% of its previousspecies diversity. Clearly, from thissmall sample of survivors, life didrecover, and we see the evidence aroundus now.

One way to look at the rebound is tofind out how long it took for particularecosystems to recover some measure ofpre-extinction complexity. We sawearlier how the latest Permian reptiles ofsouthern Africa and Russia were

organized in complex communities, withseveral scales of herbivores,presumably feeding on different gradesand sizes of plants.6 Preying on them inturn were likewise several scales ofcarnivore, from small to medium to largeto very large. The top carnivores werethe sabre-toothed gorgonopsians, whichfed on hippopotamus-sized herbivoressuch as pareiasaurs and largerdicynodonts.

Then, after the crisis, the world wasfilled with a single reptile, thedicynodont Lystrosaurus. Never beforehad one vertebrate existed in suchnumbers and on every continent. Recalltoo that Lystrosaurus made up as muchas 95% of the fauna, all the other smaller

amphibians and reptiles being incrediblyrare. Clearly, the post-apocalypticecosystem on land was a caricature of anatural system. Not only was it horriblyunbalanced, but it seemed to be almostidentical worldwide.

The story of reptilian evolutionduring the Triassic is one of expansionand increasing complexity. Thesurviving forms such as Lystrosaurusbranched out and gave rise to newreptilian dynasties. There was somedifferentiation of faunas from continentto continent. Larger, and smaller, formsappeared. Ecosystems became morecomplex.

Typical herbivores were still thedicynodonts, descendants of the group

that had been important in the LatePermian, but which had been almostwiped out, but for the survival ofLystrosaurus and its smaller relativeMyosaurus. Another synapsid group, thecynodonts, gave rise to someherbivorous side-branches, such as thechiniquodonts and diademodonts.Important too were the rhynchosaurs, adiapsid group, distant relatives of thearchosaurs. Rhynchosaurs were 1 or 2metres long, with broad triangular-shaped skulls. The front of the skull wasproduced into a pair of curving tusk-likestructures, which may have been used inraking together plant food. Their jawswere lined with multiple rows of teetharranged in a kind of cobbled pavement,

and with a groove and blade system thatprovided precise interlocking of upperand lower jaws. Evidently their plantdiet consisted of tough stems and leavesthat required firm chopping.

These diverse herbivores werepreyed on by a new array of flesh-eaters.Among the synapsids, the cynodontsdiversified into several lines of mainlysmall- and medium-sized flesh-eaters.Dominating the upper end of the scalewere members of a new group, thearchosaurs. The archosaurs, ‘rulingreptiles’, had arisen in the Late Permian,but were rare. In the Early Triassic, theybegan their inexorable rise, diversifyingslowly, and expanding their range ofniches. By the Mid Triassic the group

had split two ways, one line evolvingtowards the crocodilians, the other to thebirds. On the crocodilian line were therauisuchians, a group of terrifying flesh-eaters that existed right to the end of theTriassic. Rauisuchians ranged in size upto 5 metres long, and their deep, longskulls were lined with fearsome teeth.Here were some effective top predators,not sabre-tooths like the Late Permiangorgonopsians, but able to deal with thelargest and slipperiest herbivores of theday.

So a measure of ecosystemcomplexity had been recovered by thebeginning of the Late Triassic, some 230million years ago, and 20 million yearsafter the end-Permian crisis. There were

small, medium and large plant-eaters,and small, medium and large flesh-eaters. Twenty million years seemsrelatively rapid for biological recoveryafter such a devastating mass extinction.But in a sense, ecological recovery wasnot yet complete. There were no trulylarge plant-eaters, and the globaldiversity of species was not yet back tolatest Permian levels. Indeed, no LateTriassic reptilian fauna anywhere in theworld was as diverse as any of the LatePermian faunas of Russia or southernAfrica.

Within the Late Triassic, at the endof the Carnian stage, about 225 millionyears ago, life on land was hit by afurther extinction event. The

dicynodonts, chiniquodonts andrhynchosaurs died out. With all three ofthe dominant herbivore groups gone,their predators were also affected. Thecauses of this end-Carnian event areuncertain, but they may relate to a dryingevent and a major floral change – thereplacement worldwide of Gondwananseed ferns by conifers.

This marks the beginning of the ageof the dinosaurs. With the dominantherbivore groups gone, something wasable to evolve to fill their place. Thefirst dinosaurs are known from theCarnian, or just before, but they weresmall insect-eaters and predators thatnipped about trying to avoid the beadygaze of the rauisuchians. Indeed, the first

Carnian dinosaurs were rare, only twoor three individuals for every 50rhynchosaurs or dicynodonts. Then, afterthe end-Carnian extinction event,dinosaurs appeared all over the place.The group evidently took its chance andradiated to fill the empty ecospace.

The few modest-sized herbivorousdinosaurs that existed in the Carnianevolved into giants like Plateosaurus, abiped-cum-quadruped, with a long neckand tail, and jaws lined with sharp-edged, plant-cutting teeth. Plateosaurushad a huge sickle-like thumb that mayhave been used in raking plants together,or in defence. Either way, somespecimens of Plateosaurus became verylarge, around 5 or 7 metres long. Other

relatives achieved lengths of 10 metresor more before the end of the Triassic.At the time, flesh-eating dinosaurs wereabundant, but most were about human-sized or smaller. The top carnivoreswere still the rauisuchians, but even theyprobably could not prey on full-grownPlateosaurus, which were just too big.

After the end of the Triassic, and theextinction of the rauisuchians and someother basal archosaur groups, thedinosaurs diversified further. Largepredatory forms appeared, progenitorso f Allosaurus and Tyrannosaurus,which were big enough to tacklePlateosaurus and its relatives. So, withthe appearance of large herbivores andnew top predators in the Early Jurassic,

perhaps one could say that pre-extinctionlevels of ecosystem complexity had atlast been achieved, some 50 millionyears after the end-Permian crisis.

Disaster taxa

Similar phenomena may be seen in therecovery of marine life.7 After the massextinction, life in the sea was sparse andmonotonous, essentially identical frompole to Equator. All the rich diversity oflife in the reefs and sea floors of the LatePermian had gone. The clear differencesbetween the species that lived in thetropical Tethys Ocean and those of the

northern Boreal Ocean had disappeared.As we saw, the survivors were arestricted assemblage of rather unusualanimals.

What is becoming clear is that all therules change after a profoundenvironmental crisis. Disaster taxaprove the point (Fig. 45). These arespecies which, for whatever reason, areable to thrive in conditions that makeother species quail. After the end-Permian crisis, the inarticulatebrachiopod Lingula flourished for abrief spell, before retiring to the wings.Lingula is sometimes called a ‘livingfossil’, since it is a genus that has beenknown for most of the past 500 millionyears, and it lives today in low-oxygen

estuarine muds. It is an inarticulatebrachiopod – not because it can’t speak(although that is doubtless true), butsince it lacks a complex hingemechanism, as seen in the more abundantarticulate brachiopods. So, after a briefburst of excitement for Lingula in theanoxic Early Triassic oceans, it fadedback into the proper obscurity that hascharacterized the rest of its long history.

45 The earliest Triassic ‘disaster’forms: the brachiopod Lingula (a),and the bivalves Claraia (b),Eumorphotis (c), Unionites (d) andPromyalina (e). These magnificentfive shelly creatures were virtuallyall that survived through the end-Permian mass extinction, and theydominated black, anoxic sea floorsworldwide.

Among bivalves, famously, a groupof four disaster taxa, Claraia,Eumorphotis, Unionites andPromyalina, are found in the black,anoxic shales everywhere. Of these, thefirst three radiated during the EarlyTriassic, producing a cluster of newspecies. They were presumablyresponding to the fact that precious littleelse was around on the sea floor at thetime, and they clearly had theadaptations necessary to survive in thelow-oxygen, low-productivity, post-extinction conditions. Overall, thediversification of bivalves was a slowprocess, with new forms appearing at aslow rate. Something like the pre-extinction diversity was only recovered

by the Mid Triassic.Other marine groups seemingly

recovered faster. The ammonoids,coiled swimming cephalopod molluscs,just squeezed through the end-Permiancrisis, with the survival of only twofamilies, the Otocerataceae andXenodiscidae. The ammonoids re-radiated in two pulses, first in the latterpart of the Early Triassic, and then in theMid to Late Triassic, when theyexceeded their pre-extinction diversitywith over 150 genera.

The ‘reef gap’ following the end-Permian extinction was one of the mostprofound pieces of evidence of majorenvironmental crisis. The rich tropicalreefs of the Late Permian had all gone,

and nothing faintly resembling a reefwas seen for 10 million years after theevent. This can hardly be a result ofpoor study or collecting. Not a singlecoral specimen, a bryozoan or any otherreef animal has been found. What wereonce huge structures, often tens orhundreds of kilometres across, anddominating many coastal strips, had goneentirely. When the first tentative reefsreassembled themselves in the MidTriassic, they were composed of amotley selection of Permian survivors, afew species of bryozoans, stony algaeand sponges.

There are various kinds of reefs. Wethink of structural reefs as typical – greatwalls of coral skeletons, often built up

over millennia, and sometimes tens ofkilometres long. But the Mid Triassicreefs were modest affairs termed ‘patchreefs’, that is, low amalgamations ofreef-like creatures forming a littlecluster on the sea bed. The scleractiniancorals were there, close relatives of thecorals that abound in tropical seas today;they were diverse, but rare. It tookanother 10 million years before thesecorals had become relatively common,in the Late Triassic, and before reefsgrew in size and complexity again. Butthey were still much smaller than in theirheyday in the Late Permian.

Reefs show it, ammonoids show itand bivalves show it. The reptiles andplants too. One of the surprising recent

discoveries about the Early Triassicrecovery of life in the sea is that therewas this apparent gap of 10 millionyears before recovery really began, inthe Mid Triassic. Evolution was ineffect suspended. Simplistically, onemight have expected the recovery tobegin at once. After all, the lands andseas had been stripped of life. Normally,plants and animals live in tight patternsof ecological harmony, kept in balanceby day-to-day interactions such ascompetition and predation. If onespecies is pulled out of the system, theothers will soon move in and take over.So why did life not begin to expand anddiversify at once?

There are three suggestions. The first

is that the delay is apparent, not real. Forwhatever reason, palaeontologists havesimply failed to collect fossils in rocksof Early Triassic age, and there was nodelay. Life was burgeoning and burstingto evolve, but the fossils either were notpreserved, or they have been missed.This is always a hard argument to refute,since all the palaeontologist can say,perhaps rather plaintively or tetchily, isthat he or she has looked damned hard,and there really is nothing there. Tens ofthousands of person hours have beenspent by competent palaeontologists,who seem to manage to find fossils inabundance elsewhere, poring throughEarly Triassic rocks. What do they find?Lystrosaurus and Claraia, and nothing

else. The longer and harder they look,and as they continue to find nothing, onehas perhaps finally to believe that thegap is real.

So if the gap is real, what was theproblem? The two suggestions are thateither conditions were so harsh thatnothing could live, or that the end-Permian crisis had been so profound thatit knocked out all normal ecological andevolutionary processes. Or maybe bothfactors were in operation.

Post-apocalyptic misery

There is little doubt that the earliest

Triassic world was grim. Oceansworldwide were at a standstill, withanoxic waters widespread anddeposition of black shales common.Pyrite was forming in these shales, andother chemical anomalies indicate thatnormal processes had ceased. The low-oxygen conditions of the deep sea floorsuggest that normal oceanic circulationhad stopped, or slowed down.Normally, there is mixing of deep, coldsea-bottom waters with the warmersurface waters. The process moves at astately pace, and full mixing may takedecades, but it happens.

If normal mixing stops, the nutrientcycling processes would be destroyed.Upwelling, cold, ocean-bottom waters

cycle organic nutrients along certaincoastal margins. The most famousexample of this is along the west coastof South America, where the nutrientfeast from the deep encourages hugeshoals of fish (and huge herds offishermen, human and avian). If organicmatter were lying undisturbed on the seabed – and it was, as witnessed by theblack carbon-rich shales that were beinglaid down in the deep oceans and basins– there could have been no nutrientcycling by upwelling.

The oceanic anoxia expanded intoshallow waters too. So shallow seas,normally teeming with life, also sufferedlow-oxygen conditions. This would havedevastated all the groups that normally

rely on abundant oxygen and nutrients,and while these horrific conditionscontinued they could not re-establishthemselves. It has been estimated thatanoxic conditions persisted in shallowsea waters for several hundred thousand,or perhaps a million years. The silent,anoxic ocean floor persisted for muchlonger, probably the full 10 millionyears of the cessation of evolution.

Conditions were no better on land.The low-oxygen conditions in the oceansaffected the atmosphere as well, andevidence comes from the ancient soilslaid down during this interval. Thefamous ‘coal gap’ of the Early and earlyMid Triassic8 was a time of sparsevegetation. The extinction crisis had

stripped the land of plant life, anderosion became rapid. The EarlyTriassic soils were sparse and showed alow diversity of plants that werespecialized in surviving in hot, acidicconditions. After 10 million years,finally, in the Mid Triassic, the first thincoal seams are found. Coal indicatesrelatively lush, usually tropical-typevegetation. It was only in the LateTriassic, some 20 or 25 million yearsafter the mass extinction, that thickercoal seams are found, something likethose that were formed in the LatePermian.

The second proposal, that life washeld in check by a breakdown in thenormal rules of ecology and evolution,

seems less likely. Of course, for theinitial thousands of years after the crisis,such effects could be imagined, but 10million years is a long time for anygroup of organisms to remain in stasis.After the crisis, many of the survivingspecies would have been present in onlysmall population sizes. Under normalconditions, each species has an effectivepopulation structure, in whichindividuals breed widely, and withmovement between populations. Thiskeeps up a good genetic ferment, andmaintains a breadth of geneticpossibilities for evolution, should thatbe required.

Low population sizes can lead toproblems. The scope of genetic variation

is severely limited. Indeed, wholeswathes of the genome have been lost,and the survival of 5 to 10% of speciesinto the earliest Triassic implies an evenmore severe curtailment of the overallnumber and variation of genes. So therules of evolution were almost certainlydifferent for a while, until populationsizes of species built up again. It is hardto believe, however, that unnaturallylow population sizes could have beenmaintained for 10 million years. Thedelay in recovery must have had more todo with the abysmal Early Triassicworld.

As if that were not enough, there wasa follow-up extinction event, towardsthe end of the Olenekian stage, some 5 or

so million years after the end-Permiancrisis. Ammonoids were again hit hard,almost disappearing entirely. Thebivalves Claraia and Eumorphotis,disaster taxa that had radiated in theEarly Triassic, disappeared. There wereextinctions on land too, with variousamphibians and reptiles, most of themso-called ‘holdover’ taxa from thePermian, finally succumbing. Some ofthe extinctions might be tied to a returnto more normal environmentalconditions, with increases inoxygenation and productivity. Or theremay have been a further crisis. Thisevent is so far little understood, partlybecause it has been recognized onlyrelatively recently.

Bouncing back

The post-crisis rebound can be definedalso in purely numerical terms. One easyway to do this is to find out when globaldiversity hit pre-extinction levels. Atfamily and generic level, it took until theEarly Cretaceous, at least 100 millionyears, for marine life to reach LatePermian levels again. The global patternof recovery at species level cannotreadily be determined, but if there is atrend of increasing time to recovery atlower taxonomic levels, it might havebeen as long as 150 million years.

After the other ‘big five’ massextinctions, the rebound phase wasfaster. The best-studied has been, of

course, the KT event. For most groups ofplants and animals, it seems that therecovery phase lasted for 10 millionyears in total. There was indeed a delaybefore recovery processes kicked in, butthe niches vacated by the dinosaurs werelargely filled 10 million years after theKT event. Birds and mammals whichsurvived the KT event famously radiatedin the Palaeocene and Eocene. Indeed,as in the post-Permian world, there wasa time when the outcomes were notclear, with new groups popping up anddying out rather more quickly thannormal.

The post-KT world witnessed giantterror birds that preyed on horses. Hadevolution proceeded a little differently,

such might be the scene today. Mammalshad been effective, but small, insect-eaters, gnawers and tree climbers in thedinosaurian world. Birds had been fish-and flesh-eaters, and some had becomeflightless. So it is perhaps not surprisingthat some giant flightless birds becametop predators in different parts of theworld, hunting the smaller mammals.They had deep, massive beaks, and ranat speed, something like an ostrich,snatching their prey from the ground,and, with a flip of the neck, swallowingthe struggling mammal just about whole.At that time, the distant ancestor ofhorses, Hyracotherium, was about thesize of a terrier dog, so it would havebeen within the dietary range of some of

the predatory birds.This sort of short-lived experiment

is typical of recoveries. The firstspecies to become established may havea good evolutionary run for their money.New species can proliferate at a fasterthan normal rate as vacated niches fillup, and as ecosystems reconstructthemselves. But some of the first-comersmay not be able to cling on to theirpositions, if, for instance, they are not aswell adapted to their roles as otherspecies that come in and take over. In theend, the giant terror birds gave way inmost places to mammalian predators, theancestors of cats and dogs. So, during apost-extinction rebound there can be aphase of rapid niche-filling, and then

comes the inevitable sorting andstabilization phase when manyextinctions happen, and ecosystemsreadjust to a pattern that may then holdsway for 50 million years or more ofrelative stability.

Panic or complacency?

Lessons from the past can be read in twoways. The ecological activist wouldemphasize how human activities aredestroying biodiversity and how thiscould turn into a cascade of death asspecies after species becomes extinct.As tropical forests are cleared and reefsare poisoned, we are losing not only

species, but whole habitats. Thepalaeontological record of massextinctions then makes grimcomparisons. We know that after a massextinction, life takes a long time torecover. The geologist may say that 10million years is a short time, butmeasured in human lifetimes, it iseffectively infinity.

Low levels of extinction can turninto high levels. Destroying species andhabitats piecemeal might lead to arunaway crisis, as seems to havehappened in the past. Once the worldbecomes locked into a spiral ofdownward decline, it is impossible tosee how any intervention by humanscould turn it back. It could be, for

example, that removing one or twospecies from an ecosystem does littledamage. The remaining species canadapt and plug the gaps. But if anotherfew species are picked off, then anotherfew, and then a few more, a point maybe reached when that ecosystem willcollapse. Better to stop destroying theenvironment before we become lockedinto such a catastrophic sequence ofevents. The natural world is complex,and consequences are oftenunpredictable.

Destruction of forests can kill oceanfish, for example. Plants take up carbondioxide during photosynthesis and pumpout oxygen. Animals require oxygen, butproduce carbon dioxide as a waste

product. There is a balance here, andthat balance could be perturbed bydestroying too much of the world’sforests. Cycling of carbon is importanttoo, as dead plants and animals areincorporated into the soil, as organiccarbon builds up in the bottoms of lakesor is washed into the sea. Nutrients fromthese sources then circulate in theoceans, providing sustenance for fishes.

A political conservative could alsoclaim justification from the fossilrecord. Such a person would note thatlife has always bounced back, even froma mass extinction as profound as the end-Permian event. Evidently, each specieslocks up a huge evolutionary potential inits genes and given the chance to explore

the extent of its full capabilities, mostspecies seem to be able to proliferateand expand into new niches. Indeed, theconservative, warming to the theme,might suggest that species that have beenkilled by human intervention wereobviously rather feeble, and a bit ofextinction is good for the moral fibre.Who needs dodos and great auksanyway?

This conservative viewpoint hasgained ground in some political circles.Bjorn Lomborg,9 a Danish statisticianand one-time green campaigner, arguesthat world resources are not running out,that forest cover across the world hasincreased and that the world’s speciesare not disappearing at an alarming rate.

His views have inevitably been greetedwith outrage by many, who claim that hehas selected narrow definitions ofnatural phenomena in order to make hiscase: farmed, temperate-climate treenurseries differ from ancient, complextropical forests. It is startling none theless that it is still difficult to makedefinitive and universally convincingstatements about the state of the naturalworld today.

Coming back to reality

Much as one might wish to accept suchreassuring claims, they are toocomplacent. Of course some life will

survive human depredations. It may becockroaches or rats, but to claim thathumans cannot drive all life to extinctionis hardly cause for congratulation.

There are lessons to be learnt fromthe past. Human activities have donemore than eliminate just one or twospecies here and there. The first Maorisin New Zealand killed all the moas,some 13 or more species of impressive,large, flightless birds, thus eliminatingan entire family of birds, theDinornithidae, some time before 1775.The Maoris also killed off other, smallerfamilies of native birds. Similarly, afterEuropeans arrived on the Hawaiianislands in 1788, 18 bird speciesdisappeared, and another 12 species

may be extinct. Of 980 species of nativeHawaiian plants, 84 have already beeneliminated, and a further 133 have wildpopulations numbering fewer than 100individuals.

The famous ‘red books’ of theInternational Union for the Conservationof Nature classify different levels ofthreat to present species. Thousands ofspecies are listed as in danger ofextinction, and the lists become longerand longer each time they are revised.Species under threat include much-lovedforms such as pandas, tigers and bluewhales. Millions of dollars are spent bygovernments and charities in order to tryto conserve such species. Specialbreeding programmes in zoos help the

effort, and sometimes – rarely – thesehuge efforts allow a species to come offthe endangered list. But at what cost?We can eliminate a species in a moment,but conservation is expensive. And ofcourse, while people will pay for therescue of the panda, the Californiacondor, even the Kerry slug, who willpay for the protection of the countlessuninteresting beetles, bugs, scorpions,frogs, snakes and tropical plants that arejust as close to extinction? And what ofthe threatened species we don’t evenknow about?

The extinction estimates quoted byAl Gore have a firm basis in fact. Theonly substantial reason to question themis that there may be levels of extinction

resistance among species. What we havementioned so far is the extinction, to alarge extent, of extinction-prone forms,endemics restricted to single islands forexample. If there is such a sliding scale,and we are busily eliminating the moreprecarious species, perhaps rates ofextinction will decline as humans tacklethe more recalcitrant species, the onesthat just refuse to give in and die.

On the other hand, humanpopulations are increasingexponentially. The time it takes for thehuman population to double keepsdiminishing. Global populations rosefrom 100 million to 200 million betweenthe time of Christ and 1500. The 400million mark was achieved by 1700, 800

million by 1800, 1600 million by 1900,3200 million by 1980, and with currentlevels at nearly 6000 million (6 billion),the doubling time is down to 25 years.At today’s levels of human population,some 40% of global productivity hasbeen sequestered for our benefit –including all humans and theirdomesticated plants and animals. Thismeans that all other species havesomehow to get by on only 60% of theoxygen and carbon (well it’s more than60% since human and domestic wastegoes into the ‘wild’ systems) that theyhad available to them in the days ofJulius Caesar.

And even though the exponential risein global human population is damped,

or slowed down, by famines and wars,the rate continues to go up. This brings apressure that might cancel out anytendency to reduction of currentextinction rates. So, argue many, thedebate about extinction-prone andextinction-resistant species is irrelevant.They’re all going to go anyway, aswealthy nations pump pollutants into theatmosphere, and poorer peoples replacenatural habitats with poor-quality farmland. There is no room for complacency.

Unanswered questions

Ironically, extinctions in the distant pastare better understood than the current

crisis. Reversing that well-worn maxim,it may be that ‘the past is the key to thepresent’. Normally, geologists andpalaeontologists bow humbly beforescientists who work on modernphenomena. To understand how riversworked in the Permo-Triassic of theKaroo Basin, the geologist seeks advicefrom geographers and geomorphologistswho study modern river systems. Tounderstand what Archaeopteryx lookedlike, the palaeontologist consults anornithologist. In extinction studies,palaeontology has (some of) theanswers.

As we saw earlier, biologists havefailed to estimate current biodiversity,with estimates ranging from 5 to 100

million species. Biologists have alsofailed to estimate current extinctionrates, and there are robust debatesaround the figures quoted at thebeginning of the chapter by Al Gore.Palaeontologists, on the other hand, cangive good estimates of extinction rates,certainly at family and generic levels,and they have relatively reliable ways ofturning those into estimates of speciesextinction rates. So we know howprofound the end-Permian crisis was andthe scale of the KT event. We know alsohow long the recovery took, since time-scales were long.

Palaeontologists, of course, becomemore hesitant when pressed about howlong any particular extinction crisis

lasted. Their weakness is short time-scales, where the error bars on ageestimates may exceed the time intervalsin question. So, no one can say whetherthe end-Permian crisis lasted for one dayor a few thousand years.

Recently, conservation biologistshave been learning some practicalmethods from palaeontologists. Theyhave realized that it is futile to try tomake a complete inventory of all specieson the Earth. If it has taken since 1758 todocument the 1.8 million named species,and if there are currently 30 millionspecies, it will clearly take another 4000years of hard work to describe and namethe remainder. If there are 100 millionspecies on Earth, the task will take

14,000 years. Of course, the politicalconservative might argue that if we drivemore and more species to extinction,these estimates come down, andgovernments can save money on thesalaries of systematists. In view of thefutility of attempting to make accuratespecies counts of world diversity,conservation biologists areexperimenting with counts of genera orfamilies, and then using rarefaction toextrapolate what these figures mean atthe level of species, just as David Raupdid when he explored the scale of theend-Permian crisis.

We have focused here on the worstcrisis of all, and it is one of a number ofprofoundly unpleasant times when life

has come to the brink. Not explored herehave been those times of apparent crisiswhen not much happened to life. Therehave, for example, been numerous vastflood basalt eruptions when nothingreally became extinct, except,presumably, those unfortunate organismsthat were caught up in the flow of thelava. There have also been manyimpacts of meteorites and comets on theEarth which have not led to extinction.Some of these impacts were indeednearly as large as that which caused theChicxulub crater, and yet thepalaeontological record passes them bywith not a hint of any elevated extinctionrates. These are mysteries, yet to beresolved.

Palaeontologists and geologists arebeginning to identify common aspects ofmass extinctions. They are looking forspecies that are more prone to extinctionthan others. They are now taking moreinterest in the post-extinction recoveryperiod. The scope and timing of therebound seems to depend on themagnitude of the preceding extinction.But, perhaps unexpectedly, there is oftena long lag period when evolution seemsto have been suspended, and this wasparticularly true for the post-Permianrecovery. It is important to identifydisaster species, the forms that are ableto radiate soon after a crisis. Whatspecial features, if any, do they have thatallow them to survive, and to re-radiate

before anything else can? Or are theyjust lucky? What of the species thatfinally rebuild normal ecosystems, andreplace the disaster species?

I have tried to show in this book, byweaving history and science, howarguments are often re-run generationafter generation. Scientists are evidentlyhuman too. They can be prejudiced, theycan be scared or constrained.Catastrophic extinctions in thegeological past is a beautiful example ofan idea that was presented in the 1820s,that was firmly crushed by Lyell in the1830s, and could only raise itsdangerous head again in the 1980s. Ittook 150 years for geologists to dare toaccept the obvious, that there truly have

been mass extinctions in the past and thatstructures on the Earth’s surface thatlook like impact craters actually areimpact craters.

To have lived through the tail-end ofthis switch-over has been fascinating. Iwas taught by anti-catastrophists, but Inow preach asteroids and massextinctions to my students. Even morerapid has been the accumulation ofknowledge about the end-Permian event.Everything changed between 1992 and2001. In that time, the event focuseddown from 10 million years to a fewthousand years. The Siberian Trapsflowed into view as the main culprit.Gas hydrates, undreamt of before the1970s, suddenly became the answer to

abrupt climate changes in the past, andperhaps crucial as part of a runawaygreenhouse model for the end-Permianenvironmental crash. Long drawn-outpost-apocalyptic anoxia is everywherein the Early Triassic.

Such rapid accumulation ofknowledge and ideas has its risks. Thisbook may thereby have a short shelf-lifeas scientists disprove all the wild-eyedtheories that were published in the lastyears of the twentieth century. Fin desiècle excess, they will claim. Perhapsnot, though.

As one grows older, one realizeshow little one knows: ‘the more youlearn, the more ignorant you become’.The joy of being a scientist is to

discover this. When I was beginning mycareer, I felt that scientific research wasa line of work that led to ever greatercomplexity. As one accumulatedinformation about how the Earth works,all the simple questions would beanswered. Then the questions wouldhave to become more intricate andharder to solve.

But the unanswered questions are asbig and as simple as you could wish for(although the answers may be sointricate as to be unattainable). Howdiverse is life? How does the worldreact to human intervention? What willhappen in the next 100 years? Where didlife come from? How resilient is life tocrisis?

As Georges Cuvier wrote in 1812, atthe very birth of geology, in hisRevolutions de la globe:10

The ancient history of the earth, the ultimategoal towards which all this research is leading,is in itself one of the most fascinating subjectson which the attention of enlightened personscan be fixed. If they take an interest infollowing, in the infancy of our own species,the almost erased traces of so many extinctnations, they will doubtless find it also ingathering, in the darkness of the earth’sinfancy, the traces of revolutions previous tothe existence of every nation.

GLOSSARY

Ammonites: extinct cephalopods with coiledshells which lived abundantly as free-swimmingcarnivores in Jurassic and Cretaceous seas.

Anapsid: a reptile with no temporal openingsin the skull; includes turtles and several extinctgroups.

Angiosperm: a flowering plant.

Anoxia: lack of oxygen. The term usuallyrefers to conditions on the sea bed whenoxygen is virtually, or completely, absent.

Belemnites: extinct cephalopods withstraight, bullet-shaped internal guards.

Biodiversity: the diversity of life.

Biostratigraphy: the use of fossils toestablish the relative sequence of rocks, and tomatch rocks of similar age in different parts ofthe world.

Brachiopod: a ‘lamp shell’; member of thePhylum Brachiopoda, a group of two-shelledanimals that filter feed and generally liveattached to the sea floor. Formerly a dominantgroup, now rare.

Cephalopod: a member of the Cephalopoda,the group of molluscs that includes octopus,squid and cuttlefish, as well as the pearlynautilus and the extinct ammonites andbelemnites. All have big goggly eyes and theyare reputed to be more intelligent than otherinvertebrates.

Clade: a taxonomic group that had a singleancestor and which includes all descendants ofthat ancestor. Clades can be species, genera,

families, orders, phyla or any other formallynamed group.

Cladistics: the method of establishingphylogenies by identifying clades on the basisof unique shared characters.

Conodont: a toothed phosphatic fossil,probably the jaw parts of an extinct group ofvertebrates.

Correlation: matching rock ages from placeto place within a local basin of deposition, orglobally.

Diapsid: a reptile with two temporal openingsin the skull; includes lizards, snakes,crocodiles, birds and several extinct groups.

Dicynodont: ‘two-canine-teeth’; a member ofthe group of mammal-like reptiles that weredominant plant-eaters in the Late Permian, and

survived, in the form of Lystrosaurus, throughthe end-Permian mass extinction, to flourishagain in the Triassic.

Dinocephalian: member of an early group ofmammal-like reptiles that included plant- andflesh-eaters, and that dominated in the earlierparts of the Late Permian.

Diversity: the richness of life, usuallymeasured as the number of species, genera orfamilies, in a restricted area, or worldwide.Other definitions can include measures ofabundance, ecological range or genetic range.

Echinoderm: ‘leathery skin’, member of thePhylum Echinodermata, which includes all theinvertebrates with calcite plates and a five-rayed anatomy – the starfish, sea urchins andsea cucumbers.

Ecosystem: the physical and biological

components that make up a life environment;the environment plus the organisms that live init.

Endemic: local in distribution.

Eocene: the epoch lasting from 55–34 myr.

Fauna: the sum total of animals in a definedplace and time.

Flora: the sum total of plants in a definedplace and time.

Foraminifera: the group of microscopic,single-celled animals with calcareous shellsthat lived on the seabed and in the plankton, andwhich are useful in dating rocks.

Fusilinids: a group of radiolarians, with adelicate skeleton made from silica;microscopic animals that mainly float in the

plankton.

Gastropod: a snail, conch or limpet; memberof the group of molluscs that are equipped witha single valve (‘shell’), and the shell is oftencoiled or spired.

Goniatite: member of a cephalopod groupthat dominated oceans through the Devonian toPermian, often with a coiled shell.

Gymnosperm: a member of the plant groupthat includes conifers and smaller groups suchas cycads and ginkgos.

Igneous rocks: rocks that formed from moltenmagma, either deep within the Earth’s crust, oron the surface. Surface igneous rocks typicallyform from volcanic lavas.

Magnetostratigraphy: the division ofgeologic time using measurements of the

Earth’s magnetization as preserved in the rocks.Numerous phases of ‘normal’ (as at present)and ‘reversed’ (poles reversed) magnetizationmay be distinguished, and used for correlation.

Metamorphic rocks: sedimentary or igneousrocks that have been modified by heat orpressure after deep burial in the Earth’s crust.

Miocene: the epoch lasting from 24–5 myr.

Niche: the role and ecological attributes of aspecies, its diet, its interactions with otherspecies and its range of environmentalrequirements.

Ostracod: a small arthropod, a relative ofcrabs and shrimps, that lives inside a two-valvedshell, and swims on its back, filter feedingusing its legs.

Pecten: a bivalve mollusc with a triangular-

shaped shell, often bearing distinct radiatingridges; the symbol of the Shell PetroleumCompany.

Pelycosaurs: basal mammal-like reptiles thatwere present in the Late Carboniferous andEarly Permian; includes famously Dimetrodonwith the sail on its back.

Phylogeny: the shape of evolutionary trees, oran individual evolutionary tree.

Radiation: expansion or diversification; a timewhen a group speciates and splits rapidly,perhaps as a result of some new opportunity(e.g. after a mass extinction) or as a result ofsome new adaptation (e.g. the explosion ofmice and their relatives in the past 15 myr.).

Radiolaria: silica-shelled microscopicplanktonic animals; includes the fusilinids.

Radiometric dating: the establishment ofexact ages by study of unstable radioactiveelements and a comparison of proportions ofparent and daughter elements where the half-life of breakdown is known.

Sedimentary rocks: rocks that have formedfrom sediments, such as mud, sand or limemud, forming sandstone, mudstone orlimestone, respectively.

Stratigraphy: study of the sequence and datingof rocks.

Synapsid: a reptile with only a lower temporalopening; includes mammal-like reptiles andmammals.

Taphonomy: the processes that affect fossilsbetween the death of an organism and thediscovery of the fossil. Taphonomic processesinclude predation and scavenging damage to a

carcass, decay processes, compression anddistortion during burial, and physical andchemical alterations during and after burial.

Taxonomy: the science of classification; thedivision of the diversity of life into manageablesubgroupings, from species upwards.

Tetrapod: ‘four-footer’; Tetrapodacollectively includes all the land vertebrates –amphibians, reptiles, birds and mammals.

Therapsids: the more derived mammal-likereptiles, following after the pelycosaurs;includes all the Late Permian and Triassicsynapsids.

Therocephalians: a group of small, insect- andflesh-eating mammal-like reptiles, particularlyfrom the Late Permian and Early Triassic.

Trilobites: extinct marine arthropods, with

three-lobed bodies and many pairs of limbs,which dominated early Palaeozoic faunas ascarnivores and detritivores.

NOTESPrologue1. The quotations are from the web site

http://atlantisonline.smfforfree2.com/index.php?topic=1757.0

1. Antediluvian Sauria1. The quotation is from Owen (1845b, p.

638).2. There are many accounts of the evolution

of amphibians and reptiles, including R. L.Carroll (1988) and Benton (1997).

3. The original descriptions of the Yorkshirecrocodile are to be found in Chapman(1758) and Wooler (1758). The tale istold evocatively by Osborne (1998).

4. There is a great deal of documentaryevidence about the American mastodondebate, on both sides of the Atlantic. The

story is told in considerable detail byGreene (1961, chapter 4), and in morecompact form by Rudwick (1976), Durantand Rolfe (1984), and Buffetaut (1987).

5. There are many biographies of Cuvier. Oneof the best is by Outram (1984), whoexplores Cuvier’s life, and the mix ofscientific endeavour and administrativepower throughout his career. Rudwick(1997) focuses on his scientific work, andhe presents authoritative translations ofmany key works, with incisivecommentary.

6. The early history of discoveries ofPermian and Triassic fossil amphibiansand reptiles in Russia is told by Ochev andSurkov (2000).

7. Excellent accounts of the early discoveriesof dinosaurs in England may be found inColbert (1968) and Cadbury (2000).

8. Biographical details about Richard Owenand his work on dinosaurs may be found inbooks by Rupke (1994) and Cadbury(2000).

9. The published report by Owen on theterrestrial reptiles is Owen (1842); heestablished the Dinosauria on page 103.

10. The history of studies of the Permo-Triassic amphibians and reptiles of Russiais summarized by Ochev and Surkov(2000), and they give fuller references tothe works of Kutorga, von Qualen, vonWaldheim and von Eichwald.

2. Murchison Names the Permian1. There is no single, comprehensive

biography of Roderick Impey Murchison.The Life by Geikie (1875) and the morerecent biography by Stafford (1989) covermost aspects of his life, but the first isperhaps over-generous to its subject, the

second under-generous. Full accounts ofMurchison’s work and his character areoffered in a series of recent books thatexplore some of the particularcontroversies in which he was engaged,for example, Rudwick (1985) on the greatDevonian controversy, Secord (1986) onthe Cambrian-Silurian dispute, Oldroyd(1990) on the Highlands controversy, andCollie and Diemer (1995) on Murchison’swork in northeast Scotland.

2. Murchison named the Permian system inthe paper published in 1841, and hediscussed this further in his PresidentialAddress to the Geological Society ofLondon in February 1842 (Murchison1841a, b, 1842a, b).

3. Murchison established the Silurian Systemin his book The Silurian System(Murchison 1839).

4. Murchison and Sedgwick established theDevonian System formally in a short paperin 1839 (Sedgwick and Murchison 1839).

5. Murchison’s quoted comments on thePermian system, and on Herr Ermann’ssupposed slur on his originality, comefrom Murchison (1842b, pp. 665–666).

6. More detail of the Geological Society ofLondon, its purpose and structure, aregiven by Morrell and Thackray (1981) andRudwick (1985).

7. The quotation by Murchison about hisviews on the division of the geologicaltime-scale is in his Presidential Addressof 1842 (Murchison 1842b, pp. 648–649). Italics are as in the original.

8. The history of the division of thegeological time-scale, including WilliamSmith’s contribution, has been told manytimes by, for example, Zittel (1901),

Hallam (1983) and Rudwick (1976,1985). William Smith’s story is tolddelightfully by Winchester (2001).

9. The work by John Phillips is outlined byZittel (1901), Rudwick (1985) and by D.H. Erwin (1993). Phillips published hisstratigraphic ideas in many places,including Phillips (1838, 1840a, b, 1841),and his famous diagram of the diversity oflife through time is from Phillips (1860).

10. Sedgwick and Murchison described theirgeologizing trip to Belgium and Germanyin Sedgwick and Murchison (1840).

11. Published accounts of the first trip toRussia by Murchison are given inMurchison and Verneuil (1841a, b).

12. The earliest account of Russian geologywritten in English was by Strangways(1822).

13. The second trip to Russia was described by

Murchison (1841a, b, 1842a, b), byMurchison and Verneuil (1842), and byMurchison et al. (1842) on the UralMountains.

14. The justification for the biostratigraphicdistinctiveness of the Permian is given byMurchison (1841a, p. 419).

15. British Museum (Natural History),Richard Owen correspondence, 20November 1841.

16. Full details of the background to thepublication of the Geology of Russia(Murchison et al., 1845), the geologicalproblems, printing costs and negotiationswith the Russian government, are given byThackray (1978).

3. The Death of Catastrophism1. Lyell’s textbook is Lyell (1830–33). The

quotation is from volume 1, p. 123.

2. The catastrophism vs. uniformitarianismdebate is outlined by Hallam (1983,chapter 2) from a modern viewpoint.

3. The 1810 report by Cuvier is quoted in fullin a new translation by Rudwick (1997, pp.115–126). The two quotations are frompages 124 and 126 respectively.

4. The Ossemens fossiles was published asCuvier (1812), as a much revised editionin 1821–24, and reissued as a third, onlyslightly revised edition in 1825. Thequotations are from Rudwick’s (1997,chapter 15) translation, from pages 188–189 and 206–207 (Cuvier’s sections 5 and27 respectively, in his 1812 edition). Thispreliminary discourse was also issued as aseparate publication in four editions inEnglish from 1817 to 1827, in German,and in a separate French (third) edition in1826.

5. By 1842 Murchison would have hated tohave been classed as a Cuvierian, and yethe clearly was. In his Presidential Address,he argued for the primacy of vertebratesover invertebrates in stratigraphy in hisdispute with John Phillips. That seemsvery odd to a modern geologist, whowould generally hold quite the oppositeopinion. But Cuvier, himself a vertebratepalaeontologist, supplied lengthyarguments in favour of the importance offossil vertebrates, and the ways in whichthey would be more subject tocatastrophes than marine invertebrates. Ofcourse Murchison, as a young man in themid-1820s, was seeking to inculcatehimself in everything geological, and hewould surely have turned first to Cuvierfor inspiration: his Recherches sur lesossemens fossiles had just been issued ina second (1821–24) and third (1825)

edition, which Murchison might well haveread in the original, or in one of the newEnglish translations which had also justappeared. This was before Lyell had begunto publish his Principles in 1830,essentially a counterblast to Cuvier.Murchison might have sought to deny itlater, but the early training via MonsieurCuvier had stuck in the back of his mind,despite all efforts to expunge it. And,indeed, why expunge it?

6. The quotation is from Cuvier (1825, vol. 1,pp. 8–9), as translated by Gillispie (1951).

7. The comment on Lyell’s advocacy is fromSedgwick (1831). The strategy of Lyell’sPrinciples is discussed in detail byRudwick (1969), Gould (1987, chapter 4)and Hallam (1983, chapter 2).

8. Fitton’s comments on advocacy are in areview by Fitton (1839).

9. Lyell’s mixing of four meanings of‘uniformity’ has been exposed in detail byseveral historians of science, includingRudwick (1969, 1976, chapter 4), Hallam(1983, chapter 2) and Gould (1987,chapter 4). Astonishingly, most historianshad gone along with Lyell’s confusions upto that time, and this had cemented Lyell’sown presentation of the history of geologyas a tension between catastrophists, whowere in all regards misguided, anduniformitarians, who were entirelysensible and correct.

10. See chapters in Rudwick (1997) for fulltranslations, and commentaries, onCuvier’s views about geology.

11. Lyell’s non-progressionism is discussedby Bowler (1976). Benton (1982) showshow late Lyell continued his campaign. Inthe 1850s, he was eagerly seeking fieldevidence for his views, and documented

supposed Devonian lizards from Scotlandand supposed Silurian tetrapod tracks fromNorth America in his Elements ofGeology (1838).

12. Professor Ichthyosaurus was reinterpretedas an attack on Lyell by Rudwick (1975),and the case is further discussed by Gould(1987, chapter 4).

13. The quotations from unpublishednotebooks by J. D. Forbes are given byRudwick (1985, p. 74).

14. The letters are quoted by Geikie (1875,vol. 2, pp. 119–121). The comments on‘Lyell’s quietude’ were to Mr J. P. Martin.Murchison offered strong opposition, inletters, to Lyell’s championing of theSilurian vertebrate tracks from NorthAmerica, and the tracks and lizard skeletonfrom the supposed Old Red Sandstone ofElgin (see note 11). To Lyell he asserted,

‘Hitherto, we know absolutely nothing ofland animals in the Silurian world’. On thesupposed Devonian lizard from Elgin,Murchison wrote to Sedgwick in 1851,‘And he is to wag his tail next meeting, tothe infinite delight of Lyell, who isinebriate with joy … I hold … that suchproofs of the inhabitants of the land andfresh water of those early days can have noinfluence in changing our generalargument founded on marine succession.’

4. The Concept That Dared Not Speak itsName1. Bob Carroll’s statement about the apparent

absence of an end-Permian massextinction of vertebrates is in R. L. Carroll(1988, p. 589).

2. He later accepted the reality of the end-Permian mass extinction amongvertebrates (R. L. Carroll, 1997, p. 383).

3. Owen’s speculations about Mesozoicconditions and about the extinction of thedinosaurs are in Owen (1842, p. 203).

4. The quotations are from the third editionof the Origin of species (Darwin, 1861, p.348).

5. Huxley’s account of dinosaurian diversityis Huxley (1870).

6. Marsh’s reviews of dinosaurian diversityare Marsh (1882) and (1895).

7. Buckland’s Reliquiae Diluvianae (‘Relicsof the Flood’; 1822) gives a full accountof the findings in Kirkdale Cavern, and ofhis view that the Pleistocene extinction ofmammals had been caused by the biblicalFlood. Victorian views of Pleistoceneextinctions are discussed in detail byGrayson (1984).

8. The quotation is from Agassiz’s mostfamous work on the ice age (Agassiz

1840, p. 314; translation by the author).9. Boucher de Perthes published numerous

accounts of human remains associatedwith giant Pleistocene mammals fromFrance (Boucher de Perthes 1847, 1857,1864). Full details are given in the articleby Grayson (1984).

10. Orthogenesis (‘straight generation’) andfinalism were versions of a viewpoint thatsaw evolution as directed along pre-ordained paths. So, much of evolutioncould be said to have been directedtowards the origin of human beings, as thecrowning achievement. Bowler (1983)gives a detailed account of the history ofsuch thought in late Victorian times and inthe early twentieth century.

11. The quotations are from Woodward (1898,pp. 213 and 418).

12. Arthur Smith Woodward’s address to the

British Association is Woodward (1910).13. The quotation is from Loomis (1905, p.

842).14. The quotations are from Nopcsa (1911, p.

148) and Nopcsa (1917, p. 345).15. Matthew’s ideas about dinosaurian

extinction are given in Matthew (1921).16. The other suggestions about dinosaurian

extinction, from the 1920s, were climaticcooling (Jakovlev, 1922), high diseaselevels (Moodie, 1923), mammals eatingdinosaur eggs (Wieland, 1925) andvolcanic eruptions (Müller, 1928).

17. The first major review of the extinction ofthe dinosaurs was presented by Audova(1929).

18. The two surveys of theories for theextinction of the dinosaurs are Jepsen(1964) and Benton (1990).

19. Will Cuppy, renowned American

humourist, gave his view of racial senilityin his book How to Become Extinct in1964.

20. The quotation comes from page 35 of theEnglish translation of Schindewolf’sclassic book. Grundfragen derPaläontologie, published in 1950. Thetranslation, Basic Questions inPaleontology was published in 1993, longafter Schindewolf’s death, and more as anhistorical curiosity than as a serious text.The fact that Schindewolf publishedexclusively in German, and in the Germanjournals, and that his book was not earliertranslated into English, confirms howdivided the English-speaking and German-speaking palaeontologists were duringmuch of the twentieth century. This haschanged now, with Germanpalaeontologists regularly writing inEnglish, as well as in German, and placing

their articles in international journals.21. Camp (1952) and Watson (1957) doubted

the reality of the end-Permian extinctionevent among amphibians and reptiles.Schindewolf (1958) argued stronglyagainst their viewpoints, and gave a fullerriposte in his 1963 ‘Neokatastrophismus?’paper.

5. Impact!1. The epochal paper, which presented the

first fully worked-out model fordinosaurian extinction by impact was L.W. Alvarez et al. (1980).

2. Numerous accounts of the KT debatessince 1980 have been published. W.Alvarez (1997) is the best insider’s viewof the work by Luis Alvarez and his team.Other books about the KT impact and thedebates include Raup (1986), whichpresents a strong palaeontological case

for impact, Archibald (1996), whichpresents the opposite case – for cautionabout accepting a simple, instantaneousimpact model – and Glen (1994), whichcontains chapters by a number of authorson the scientific disputes.

3. Ken Hsü (1994) gives a personal memoirof the shift from dogmatic anti-catastrophism in the United States to theacceptance of the possibility of majorimpacts on the Earth during the twentiethcentury.

4. The story of the Ries crater, includingShoemaker’s involvement is told in manyplaces, including guidebooks by Kavasch(1986) and Chao et al. (1978). Stöfflerand Ostertag (1983) provide a moretechnical account of the crater and theevidence for impact. The critical firstdemonstration that the Ries structure wasan impact crater is Shoemaker and Chao

(1961).5. The gradualist ecological succession

model for the extinction of dinosaurs, andother animals and plants, at the KTboundary, was summarized by Van Valen(1984), Sloan et al. (1986), Officer et al.(1987) and Hallam (1987).

6. The distinction between ‘hard science’impact supporters and ‘soft science’gradualists was noted by Raup (1986, p.212).

7. The quotation is from L. W. Alvarez (1983,p. 632).

8. The quotation is from Jastrow (1983, p.152).

9. The comment is from Van Valen (1984, p.122).

10. The quotation is from L. W. Alvarez(1983, p. 67).

11. These further quotations are from L. W.

Alvarez (1983), pages 638, 640 and 629respectively.

12. News reports about some of the more uglyaspects of the KT impact debate includeBrowne (1985, 1988).

13. The quotation by Robert Bakker is given byRaup (1986, pp. 104–105).

14. A sociological review of styles ofargumentation and clashes of differentbranches of science in the KT debate isgiven by Clemens (1986, 1994). See alsoother papers in Glen (1994) forcommentary on how the KT debatesevolved.

15. The supernova theory for the demise of thedinosaurs was championed in particular byDale Russell, then of the Royal OntarioMuseum in Toronto, Canada, and hiscolleagues (Terry and Tucker, 1968;Russell and Tucker, 1971; Béland et al.,

1977).16. Later papers that further explain the

consequences of impact are W. Alvarez etal. (1984a, b). The quotation is from W.Alvarez et al. (1984a, p. 1135).

17. The other articles which appeared at thesame time were Smit and Hertogen(1980), Hsu (1980) and Ganapathy(1980).

18. Informed, contemporary news coverage ofthe KT debate was given by Lewin (1983,1985a, b), Maddox (1985), Hoffman(1985) and Browne (1985, 1988).

19. The discovery of the Chicxulub crater wasreported by Hildebrand et al. (1991).

6. Diversity, Extinction and MassExtinction1. The quotation is from Cuvier (1812), as

translated by Rudwick (1997, 190).

2. D. H. Erwin’s (1993) book is a detailedaccount of the end-Permian extinctionevent.

3. The word ‘biodiversity’ was introduced inthe report by Wilson and Peter (1988) onthreats to the global diversity of life.

4. The two fundamental works in theclassification of plants and animals areLinnaeus (1753) and Linnaeus (1758)respectively.

5. Darwin (1859) is, of course, his famousOrigin of species, the starting point of somany fields of biology and palaeobiology.The basic characters of life are reviewedin all general biology textbooks. Recentreviews of the beginnings of life, and lateradditions of complexity, are MaynardSmith and Szathmáry (1995), Knoll andBambach (2000) and S. B. Carroll (2001).

6. Plots of diversification of life in the sea

were first produced in the 1960s and1970s. The most famous examples stemfrom the work by Jack Sepkoski, Jr. (e.g.Sepkoski, 1984). I have produced plots ofdiversification of life in the sea and onland (Benton, 1995) based on acomprehensive data compilation producedby 100 experts (Benton, 1993).

7. The different mass extinction events arenot all explored in detail here. There aremany books about mass extinctions, andthe best current one is by Hallam andWignall (1997). It gives a full account ofeach event, what happened and why, or atleast a selection of the hypotheses foreach event.

8. The proposed statistical test for massextinctions was presented by Raup andSepkoski (1982).

9. The statistical criticism of Raup and

Sepkoski’s method was that a regressionline can only be drawn if it is assumed thatthe data are a normal population ofindependent points. Techniques likeregression analysis had been developed bybiologists and psychologists who used themethods to sort out cause and effect inpopulations of plants, animals or humans.It was stretching the validity of theapproach to use it on fixed measuresthrough time. First, the extinction ratemeasures are not independent of eachother: clearly, they occur in a sequence,and they affect a single evolving system,so there is strong time-linkage betweenadjacent rate measures. Secondly, it is notpossible to have a complete distribution,as with a population of plants, animals andhumans. In the extinction rate case, part ofthe distribution of values, for negativeextinction rates, cannot be sampled.

10. The sampling studies of dinosaurextinction in the Hell Creek Formation ofMontana, USA, are described by Archibaldand Bryant (1990) and Sheehan et al.(1991), as well as in later papers.Archibald (1996) is an excellent bookabout the fossil record around the KTmass extinction.

11. Our study of perceptions of the fossilrecord over the past 100 years ispresented in Maxwell and Benton (1990).The comparison of changes in knowledgeof the marine fossil record over 10 yearsis given by Sepkoski (1993).

12. Peter Ward’s comparisons of hisunderstanding of ammonite extinctions inthe KT boundary sections of northernSpain are summarized in Ward (1990).

13. Cladistics was enunciated first in Englishby Hennig (1966), and there are now many

accounts of the methods. Forey et al.(1998) is a good primer of the basics.

14. The nucleic acids are deoxyribose nucleicacid, abbreviated as DNA, and ribosenucleic acid, abbreviated as RNA. DNA isthe key coding material within thechromosomes within the nucleus of cells,and the various forms of RNA transferinformation during cell division and duringprotein synthesis. Any recent basicbiology book will give fuller details. Themethods of molecular phylogenyreconstruction are outlined in Forey et al.(1998) and in Hillis et al. (1996).

15. The first comparisons of phylogenies andstratigraphy were made by Norell andNovacek (1992). They looked at 25phylogenies of mammals and found thatthree-quarters of them showed goodcongruence.

16. Benton et al. (2000a) considered 1000phylogenies, cladistic and molecular, of awide range of organisms, from algae tomammals, and found good agreement. Inaddition, when the phylogenies weresorted into those with branching occurringat different points in the geological past,they showed constant levels of agreementthrough geological time.

17. Jablonski and Raup (1995) sampled a hugedatabase on marine bivalves andgastropods across the KT boundary,comparing victims with survivors. Therewas no selectivity on the basis of habitat,size, diet or breeding habits. Butgeographically widespread genera weremore likely to survive than those withrestricted geographic ranges.

18. The role of geographic distribution onsurvival through the KT event was assessedby Raup and Jablonski (1993).

Comparisons of bivalves with temperateand tropical distributions did not confirmthat tropical forms were more likely to goextinct than temperate, which had been theexpected result.

7. Homing in on the Event1. The quotation is from Teichert (1990, p.

231).2. D. H. Erwin (1993, p. 226) gives the

estimate of 3–8 million years for the end-Permian event.

3. The reduction in timing of the end-Permianevent to less than a million years is givenby Bowring et al. (1998).

4. The further revision downwards, to lessthan 500,000 years, or even instantaneous,is given by Jin et al. (2000).

5. Murchison’s ‘reduced’ succession of thePermian is discussed in his big book

describing his adventures in Russia(Murchison et al., 1845). Detailedevidence in support of this interpretationis given by Dunbar (1940) and Harland etal. (1982, p. 23).

6. Karpinskiy named the Artinskian stage in1874, and gave a fuller account of it in1889, with full descriptions of the fossils,and evidence for their Permian, notCarboniferous, affinities. It is named afterthe town Arti (sometimes spelledArtinsk), on the western side of the UralMountains.

7. Ruzhentsev (1950) subdivided theSakmarian into a lower substage, which hecalled the Asselian, and an upperSakmarian substage. The two were laterraised to full stage status. The Sakmarianwas named after the Sakmara, a tributaryof the Ural River in the southern UralMountains. The Asselian is named after

the Assel River. The other Permian stagesare also named after places in Russia:Kungurian (Kungur, a town near Perm),Ufimian (Ufa, a city in the Urals),Kazanian (Kazan, a city on the banks of theVolga), Tatarian (after the Tatars ofTatarstan, a region of Asiatic Russia).

8. The 1937 conference is described byWilliams (1938) and a detailed overviewof the field trip and the new view ofRussian Permian stratigraphy by Dunbar(1940).

9. Dunbar’s remark comes in theacknowledgments section of his paper(Dunbar, 1940, p. 280). The formalizedprinciples of stratigraphy are given indetail in Salvador (1994). Good accountsof stratigraphy may be found in textbookssuch as Stanley (1986) and Prothero(1990).

10. The Permo-Triassic sections in northernItaly have been described by Wignall andHallam (1992), D. H. Erwin (1993, pp.52–57) and Hallam and Wignall (1997, pp.118–119).

11. Kummel and Teichert (1966, 1973)describe their visits to Permo-Triassicboundary sections around the world.

12. Restudy of the conodonts from the SaltRange is reported by Wignall et al.(1996), where conodonts in the KathwaiMember, previously said to be indicativeof Early Triassic age, are reassigned to thelatest Permian.

13. Doug Erwin commented on thecorrespondence of important Permo-Triassic boundary sections in Asia withcurrent political turmoil in his book (D. H.Erwin, 1993, p. 63).

14. The establishment of the base of the

Triassic was announced by theInternational Union of GeologicalSciences in 2000. Further details may befound in Yang et al. (1995).

15. The sedimentology of the Meishan Permo-Triassic boundary section is described byWignall and Hallam (1993), andsummarized by Hallam and Wignall (1997,pp. 120–122). Further details are given byYang et al. (1995), Jin et al. (2000), andLai et al. (2001).

16. The first radiometric dates for theMeishan boundary clays were done byClaoue-Long et al. (1991) using the206Pb/235U isotopic decay series.

17. The first argon-argon date for the Permo-Triassic boundary was reported by Renneet al. (1995).

18. Detailed dating studies on the Meishansection were done by Bowring et al.

(1998), with later corrections by Mundilet al. (2001).

19. Details of the distribution of fossilsacross the Permo-Triassic boundary in theMeishan section are given by Jin et al.(2000).

20. The Signor-Lipps effect is named after theaccount of necessary incompleteness ofthe fossil record by Signor and Lipps(1982).

21. Wignall and Hallam (1993, p. 231)comment on the similarity of the Hushansection to Meishan.

22. The Greenland Permo-Triassic successionis described by Teichert and Kummel(1976), where they present their‘armoured mudballs’ hypothesis. Hallamand Wignall (1997, p. 117) said ‘armouredmudballs’ to that theory. Dating evidenceis given by Wignall et al. (1996).

23. Twitchett et al. (2001) and Looy et al.(2001) describe the Greenland Permo-Triassic sections in some detail.

8. Life’s Biggest Challenge1. D. H. Erwin (1993) used the term ‘mother

of all mass extinctions’ as a chapter title.The quotation is from D. H. Erwin (1994,p. 231).

2. The statistical analysis of fusulineextinctions is Rampino and Adler (1998).

3. The quotation is from Darwin’s ‘bigspecies book’, his unpublished longerversion of the Origin (Stauffer 1975, p.208).

4. The reanalysis of the fates of brachiopodsand bivalves is in Gould and Calloway(1980). Other information on the fate ofdifferent groups of shellfish during theend-Permian crisis is in D. H. Erwin

(1993) and Hallam and Wignall (1997).5. A superb account of trilobites is given by

Fortey (2000). Extinctions among fossilarthropods are discussed by Briggs et al.(1988).

6. Hallam and Wignall (1997, pp. 98–110)offer a detailed overview of the end-Permian extinctions, group by group.

7. Raup (1979) presented the first simplerarefaction analysis that a loss of 50% offamilies is equivalent to the loss of 90%of species. Criticisms were made byHoffman (1986) and McKinney (1995),who claimed that this was an overestimate,and that true species-level extinctionswere much lower.

9. A Tale of Two Continents1. The story of Andrew Bain and his fossil

collecting is told in his own personal

memoir, published posthumously (Bain1896), as well as by Desmond (1982, pp.195–198) and by Buffetaut (1987, pp.176–179).

2. The first publications on the Karoo Permo-Triassic reptiles of the Karoo Basin are byBain (1845) on the geology, and by Owen(1845a) on the reptiles.

3. Robert Broom’s story is told by Watson(1952), and more briefly by Buffetaut(1987, p. 179).

4. Western accounts of the Russian CopperSandstone reptiles were given by Meyer(1866) and Owen (1876).

5. British palaeontologists who visited theRussian localities and wrote about thefossils in Victorian times, includeTwelvetrees (1880, 1882) and Seeley(1894).

6. Ivan Efremov’s contribution to Russian

palaeontology is assessed by Ochev andSurkov (2000), who also trace inconsiderable detail the work of hisstudents since the 1950s. Efremov’s keypresentations of the Permo-Triassicstratigraphic scheme are Efremov (1937,1941, 1952). His science fiction novelsstill sell widely. I was able to buy a reprintof his Tais Afinskaya (the name of amythical Athenian goddess) in 1995 in asmall shop in a South Urals’ village thathad little else on offer (some twiggyraisins, oily tins of oily fish, as well asbeautiful Western-style tins of fruitimported from Hungary, at prices no onecould afford).

7. The current stratigraphic scheme for theKaroo Permo-Triassic is summarized byRubidge (1995). James Kitching’s workon localizing Karoo vertebrate specimensis summarized in Kitching (1977). These

works summarize the faunas of theDicynodon Zone, and this is supplementedby overviews in Anderson (1999).

8. Estimates of familial losses amongterrestrial vertebrates are taken fromMaxwell (1992) and Benton (1993,1997).

9. The idea that erosion rates increased at theend of the Permian as a result of loss ofplant cover is presented by MacLeod et al.(2000) and Ward et al. (2000).Distributions of the reptiles across thePermo-Triassic boundary in the Karoo aregiven by Smith and Ward (2001).

10. Retallack (1999) gives a clear and detailedaccount of changes in soils across thePermo-Triassic boundary in Gondwana.

11. Eshet et al. (1995) describe evidence for asharp fungal spike just at the Permo-Triassic boundary. Fungi (moulds and

mushrooms) reproduce by sending outhyphae, tube-like structures that can giverise to new individuals, and form a hugeand complex underground mat. Fungi alsoproduce spores which are thrown out intothe air or water, and may spread widely.

12. Looy et al. (2001) and Twitchett et al.(2001) present their evidence about themarine and terrestrial extinction eventsfrom the Greenland sections.

13. Broom (1903) described and namedLystrosaurus and argued that it lived asemi-aquatic lifestyle. This wasquestioned seriously by King (1991b) andKing and Cluver (1991), but refuted, forthe Russian species at least, by Surkov(2002).

10. On the River Sakmara1. Andrew Newell’s account of the

sedimentology of the Russian Permo-

Triassic sections is Newell et al. (1999).The footprints in the mudflats associationwere described by Tverdokhlebov et al.(1997).

2. Valentin Tverdokhlebov has publishedmany short papers on the sediments of thePermo-Triassic of the Urals. His accountof the alluvial megafans and their sourceareas is Tverdokhlebov (1971).

3. Early accounts of the Russian Permo-Triassic amphibians and reptiles areEfremov (1940), in German, and Olson(1955, 1957), in English. Olson (1990)wrote an extremely informative account ofhis visits to Moscow, and his meetingswith Ivan Efremov, in the 1950s. Sennikov(1996) gave a brief account of theevolution of Permo-Triassic ecosystemsin Russia, as part of the joint researchprogramme between Bristol and Moscow.

4. The major outcome of the collaborationbetween Bristol and Moscow so far isBenton et al. (2000b), an edited book thatcontains chapters on the rock successionsand the major groups of amphibians andreptiles from the Permo-Triassic, as wellas younger dinosaur-bearing units ofMongolia and the Former Soviet Union.We have begun to transcribe ValentinTverdokhlebov’s card catalogue of Permo-Triassic sites, and the first part ispublished as Tverdokhlebov et al. (2002).

11. What Caused the Biggest Catastrophe ofAll Time?1. Herbert Butterfield’s book, The Whig

Interpretation of History, first publishedin 1931 and reprinted many times since, isa standard source for discussions abouthistorical method. It has been attacked anddefended equally vigorously in many

works later in the twentieth century.2. Frank Rhodes made his comment in a

review of the Permo-Triassic extinctionevent (Rhodes, 1967).

3. Quoted from D. H. Erwin (1993, p. 256).4. N. D. Newell (1952) first noted the Late

Permian marine regression and defendedthe idea in later papers, such as Newell(1967).

5. Valentine and Moores (1973) developedthe plate tectonic hypothesis for sea-levelfall and mass extinction.

6. The evidence for and against a LatePermian marine regression is presented byD. H. Erwin (1993, pp. 145–155). Hallamand Wignall (1997, pp. 132–133) castsevere doubts on the idea of regression atthis time.

7. The three extinction peaks across thePermo-Triassic boundary are shown in my

plot of diversification of continentaltetrapods (Benton, 1985), and the end-Olenekian peak is shown for bothtetrapods and for ammonoids in Benton(1986), based on global family-level andgeneric-level counts. Two peaks ofextinctions among Late Permian tetrapodswere confirmed by King (1991a) andRubidge (1995), based on species countsfrom the Karoo Basin. Stanley and Yang(1994) also pointed out the end-Capitanian extinction among marinespecies, especially as seen in the Chinesesections. Wang and Sugiyama (2000)made detailed studies of coral extinctionsthrough the Late Permian in China.Sepkoski (1996) also noted the Olenekianextinction event, sometimes called theend-Scythian event, using an alternativename for the time unit.

8. Schindewolf (1958) suggested cosmic rays

as a cause of the end-Permian event.9. The enhancement in iridium at the Permo-

Triassic boundary was announced from theMeishan section by Sun et al. (1984) andfrom another Chinese section by Xu et al.(1985). The re-analyses were done byClark et al. (1986) and Zhou and Kyte(1988), who found no unusual quantitiesof iridium.

10. The ferruginous spherules were reportedby Yin et al. (1992).

11. Kerr (1995) was first to report theshocked quartz grains from Australia andAntarctica, and Retallack et al. (1998)gave a fuller account.

12. The report of extraterrestrial noble gasesin fullerenes is Becker et al. (2001).Quoted newspaper reports are Hawkes(2001) and Hanlon (2001). Erwin andBecker were quoted by Simpson (2001).

Published critiques are Farley andMukhopadhyay (2001) and Isozaki (2001),with a response by Becker and Poreda(2001). Kaiho et al. (2001) addedgeochemical evidence for impact. Becker(2002) gives a popular account.

13. Courtillot (1990) postulated a short time-span for eruption of the Deccan Traps,exactly at the KT boundary, while Baksiand Farrar (1991) argued that theprecision of his dates was illusory. Laterwork tended to confirm Courtillot’s data.

14. Baksi and Farrar (1991) gave radiometricdates for the Siberian Traps that placedtheir eruption in the Mid Triassic. Thedates were revised to the Permo-Triassicboundary by Renne and Basu (1991), usingthe argon-argon method, and Campbell etal. (1992), using the uranium-lead method.Later redatings, using wider arrays ofspecimens, confirmed that the Siberian

basalts coincided with the Permo-Triassicboundary, and were erupted in less thanone million years (Basu et al., 1995;Renne et al., 1995). Revisions to thepostulated volume of the Siberian Trapsare given by Reichow et al. (2002).

15. The quotation is from D. H. Erwin (1993,pp. 255–256).

16. The volcanic winter hypothesis waspresented by Campbell et al. (1992). Acritique is given by Hallam and Wignall(1997, pp. 135–137).

17. Stanley (1984, 1988) argued strongly thatmass extinctions, including that at the endof the Permian, were caused by globalcooling. Hallam and Wignall (1997, pp.133–134) dispute these claims, butpresent evidence that the end-Capitanianevent might have been the result of globalcooling.

18. Hallam and Wignall (1997, pp. 137–138)summarize the evidence for globalwarming at the end of the Permian.

19. The isotope geochemistry of theGartnerkofel core was presented byHolser et al. (1989), and of theSpitsbergen section by Gruszczynski et al.(1989).

20. The palaeosol evidence for climatewarming is given by Retallack (1999), whoalso discusses changes in the floras.

21. The Spitsbergen sections are described byWignall and Twitchett (1996). Anoxicblack mudstones are reported from Italyand the United States by Wignall andHallam (1992), from Pakistan and Chinaby Wignall and Hallam (1993), and fromGreenland by Twitchett et al. (2001).

22. Hallam and Wignall (1997, chapter 5)present a detailed account of the end-

Permian crisis and their model based onoxidation of coals. Wignall (2001) bringsin the Siberian Traps, and other sources ofcarbon dioxide. Wignall and Twitchett(2002) review the whole superanoxiahypothesis.

23. The methane burp hypothesis for theglobal warming pulse 55 million years agohas been presented by Dickens et al.(1997), and enlarged on by Bains et al.(1999, 2000a, b).

24. Robert Berner outlines his climaticmodels and the end-Permian peturbationsin Berner (2002).

12. The Sixth Mass Extinction?1. Gore (1992, p. 28).2. Ehrlich and Ehrlich (1990) provide an

estimate of the current rate of speciesloss.

3. T. Erwin (1982, 1983) presented hissampling estimates of tropical arthropoddiversity.

4. Total global diversity has been reviewed bymany authors, including Wilson (1992)and May (1990, 1992). Read more inWilson (2002), his latest book. New workby Novotny et al. (2002) suggests thatErwin overestimated the level ofendemicity of his beetles; more of themare actually shared between tropical treesthan he thought. If correct, this could bringhis estimate down to 4–6 million speciesof arthropods.

5. Current extinction rates have beenestimated by Smith et al. (1993) andPimm et al. (1995). Read more in Wilson(2002), his latest book.

6. The history of tetrapods on land during theTriassic, and the origin of dinosaurs, has

been reviewed many times, by Benton(1997), for example.

7. The Early Triassic recovery phase for lifein the sea has been summarized by D. H.Erwin (1993) and Hallam and Wignall(1997). Detailed accounts have been madeby McRoberts (2001) on bivalves, Erwinand Pan (1996) on gastropods, Twitchett(1999) on the changing environments,Rodland and Bottjer (2001) on Lingula,and Twitchett et al. (2001) on marinerecovery in the Greenland sections. Analternative view, in which some initialrecovery is posited within 1 myr. of themass extinction is presented by Chen et al.(2002). They find the new brachiopodMeishanorhynchia evolving by bed 36 ofthe Meishan section, dated at 250.1 myr.

8. The post-extinction ‘coal gap’ is discussedby Retallack et al. (1996). Looy et al.(2001) describe the recovery of plants

immediately following the end-Permianmass extinction in the Greenland sections.

9. Bjorn Lomborg’s views are presented inLomborg (2001). They have beencriticized in numerous reviews, includingthree articles from the Guardian, athttp://www.guardian.co.uk/globalwarming/story/0,7369,539558,00.html

10. Quoted from Cuvier (1812), as translatedby Rudwick (1997, p. 185).

BIBLIOGRAPHY

Agassiz, L. 1840. Études sur les glaciers.Neuchâtel: Jent and Gassman.

Alvarez, L. W. 1983. Experimental evidencethat an asteroid impact led to the extinctionof many species 65 million years ago.Proceedings of the National Academy ofSciences, USA 80, pp. 627–642.

——, Alvarez, W., Asaro, F. and Michel, H. V.1980. Extraterrestrial cause for theCretaceous-Tertiary extinction –Experimental results and theoreticalimplications. Science 208, pp. 1095–1108.

Alvarez, W. 1997. T. Rex and the Crater ofDoom. Princeton, New Jersey: PrincetonUniversity Press; London: Penguin.

——, Kauffman, E. G., Surlyk, F., Alvarez, L.W., Asaro, F. and Michel, H. V. 1984a.

Impact theory of mass extinctions and theinvertebrate fossil record. Science 223, pp.1135–1141.

——, Alvarez, L. W., Asaro, F. and Michel, H.V. 1984b. The end of the Cretaceous: Sharpboundary or gradual transition? Science223, pp. 1183–1186.

Anderson, J. M. (ed.) 1999. TowardsGondwana Alive. Pretoria: GondwanaAlive Society.

Archibald, J. D. 1996. Dinosaur Extinctionand the End of an Era: What the FossilsSay. New York: Columbia UniversityPress.

—— and Bryant, L. J. 1990. DifferentialCretaceous/Tertiary extinctions ofnonmarine vertebrates: evidence fromnortheastern Montana. Special Paper ofthe Geological Society of America, 247,pp. 549–562.

Audova, A. 1929, Aussterben derMesozoischen Reptilien, Palaeobiologica2, pp. 222–245, pp. 365–401.

Bain, A. G. 1845. On the discovery of thefossil remains of bidental and other reptilesin South Africa. Quarterly Journal of theGeological Society of London 1, pp. 317–318, and Transactions of the GeologicalSociety of London, Series 2 7, pp. 53–58.

—— 1896. Reminiscences and anecdotesconcerned with the history of geology inSouth Africa, or the pursuit of knowledgeunder difficulties. Transactions of theGeological Society of South Africa 2, pp.59–75.

Bains, S., Corfield, R. and Norris, R. 1999.Mechanisms of climate warming at the endof the Paleocene. Science 285, pp. 724–727.

——, —— and —— 2000a. Structure of the

late Palaeocene carbon isotope excursion.GFF 122, pp. 19–20.

——, Norris, R. D., Corfield, R. and Faul, K. L.2000. Termination of global warmth at thePalaeocene/Eocene boundary throughproductivity feedback. Nature 407, pp.171–174.

Baksi, A. K. and Farrar, E. 1991. Ar40/Ar39

dating of the Siberian Traps, USSR –evaluation of the ages of the 2 majorextinction events relative to episodes offlood-basalt volcanism in the USSR and theDeccan Traps, India. Geology 19, pp. 461–464.

Basu, A. R., Poreda, R. J., Renne, P. R.,Teichmann, F., Vasiliev, Y. R., Sobolev, N.V. and Turrin, B. D. 1995. High-He3 plumeorigin and temporal-spatial evolution of theSiberian flood basalts. Science 269, pp.822–825.

Becker, L. 2002. Repeated blows. ScientificAmerican 286(3), pp. 62–69.

—— and Poreda, R. J. 2001. Anextraterrestrial impact at the Permian-Triassic boundary? Science 293, 2343a(U2–U3).

——, ——, Hunt, A. G, Bunch, T. E. andRampino, M. 2001. Impact event at thePermian-Triassic boundary: evidence fromextraterrestrial noble gases in fullerenes.Science 291, pp. 1530–1533.

Béland, P., Feldman, P., Foster, J., Jarzen, D.,Norris, G., Pirozynski, K., Reid, G., Roy, J.R., Russell, D. and Tucker, W. 1977.Cretaceous-Tertiary extinctions andpossible terrestrial and extraterrestrialcauses. Syllogeus 1977(12), pp. 1–162.

Benton, M. J. 1982. Progressionism in the1850s: Lyell, Owen, Mantell, and the Elginfossil reptile Leptopleuron (Telerpeton).

Archives of Natural History, 11, pp. 123–136.

—— 1985. Mass extinction among non-marinetetrapods. Nature 316, pp. 811–814.

—— 1986. More than one event in the lateTriassic mass extinction. Nature 321, pp.857–861.

—— 1990. Scientific methodologies incollision; the history of the study of theextinction of the dinosaurs. EvolutionaryBiology 24, pp. 371–400. Also available athttp://palaeo.gly.bris.ac.uk/essays/dino90.html

—— (ed.) 1993. The Fossil Record 2. London:Chapman & Hall.

—— 1995. Diversification and extinction inthe history of life. Science, 268, pp. 52–58.

—— 1997. Vertebrate Palaeontology, 2ndedition. London and New York: Chapman &Hall (reissued, Oxford: Blackwells, 2000).

—— and Harper, D. A. T. 1997. Basic

Palaeontology. London: LongmanAddison-Wesley.

——, Wills, M. A. and Hitchin, R. 2000a.Quality of the fossil record through time.Nature 403, pp. 534–538.

——, Shishkin, M. A., Unwin, D. M. andKurochkin, E. N. (eds) 2000b. The Age ofDinosaurs in Russia and Mongolia.Cambridge: Cambridge University Press.

Berner, R. A. 2002. Examination of hypothesesfor the Permo-Triassic boundary extinctionby carbon cycle modeling. Proceedings ofthe National Academy of Sciences, 99, pp.4172–4177.

Boucher de Perthes, J. 1847. Antiquitésceltiques et antédiluviennes. Mémoire surl’industrie primitive et les arts à leurorigine (Vol. 1). Paris: Treuttel and Wertz.

—— 1857. Antiquités celtiques etantédiluviennes. Mémoire sur l’industrie

primitive et les arts à leur origine (Vol.2). Paris: Treuttel and Wertz.

—— 1864. Antiquités celtiques etantédiluviennes. Mémoire sur l’industrieprimitive et les arts à leur origine (Vol.3). Paris: Treuttel and Wertz.

Bowler, P. J. 1976. Fossils and Progress.Palaeontology and the Idea ofProgressive Evolution in the NineteenthCentury. New York: Science HistoryPublications.

—— 1983, The Eclipse of Darwinism: Anti-Darwinian Evolution Theories in theDecades Around 1900. Baltimore, Md.:Johns Hopkins University Press.

Bowring, S. A., Erwin, D. H., Jin, Y. G., Martin,M. W., Davidek, K. and Wang, W. 1998.U/Pb zircon geochronology of the end-Permian mass extinction. Science 280, pp.1039–1045.

Briggs, D. E. G., Fortey, R. A. and Clarkson, E.N. K. 1988. Extinction and the fossilrecord of arthropods. In Larwood, G. P.(ed.) Extinction and Survival in the FossilRecord, pp. 171–209. SystematicsAssociation Special Volume No. 44.

Broom, R. 1903. On the remains ofLystrosaurus in the Albany Museum.Records of the Albany Museum 1, pp. 3–8.

Browne, M. W. 1985. Dinosaur experts resistmeteor extinction idea. Paleontologists saydissenters risk harm to their careers. NewYork Times 1985 (29 October), pp. 21–22.

—— 1988. Debate over dinosaur extinctiontakes an unusually rancorous turn. NewYork Times 1988 (19 January), pp. 19, 23.

Buckland, W. 1822. Reliquiae diluvianae; or,observations on the organic remainscontained in caves, fissures, and diluviangravel, and on other geological

phenomena, attesting the action of anuniversal deluge. London: John Murray.

Buffetaut, E. 1987. A Short History ofVertebrate Palaeontology. London:Croom Helm.

Butterfield, H. 1931. The Whig Interpretationof History. London: Bell. (Reprinted 1965,New York: Norton.)

Bystrov, A. P. 1957. [The pareiasaur skull.]Trudy Paleontologicheskogo Instituta ANSSSR 68, pp. 3–18.

Cadbury, D. H. 2000. The Dinosaur Hunters.London: Fourth Estate; New York: Holt.

Camp, C. L. 1952. Geological boundaries inrelation to faunal changes and diastrophism.Journal of Paleontology 26, pp. 353–358.

Campbell, I. H., Czamanske, G. K., Fedorenko,V. A., Hill, R. I. and Stepanov, V. 1992.Synchronism of the Siberian Traps and thePermian-Triassic boundary. Science 258,

pp. 1760–1763.Carroll, R. L. 1988. Vertebrate Palaeontology

and Evolution. New York: W. H. Freeman.—— 1997. Patterns and Processes of

Vertebrate Evolution. Cambridge:Cambridge University Press.

Carroll, S. B. 2001. Chance and necessity: theevolution of morphological complexity anddiversity. Nature 409, pp. 1102–1109.

Chao, E. C. Y., Hüttner, R. and Schmidt-Kaler,H. 1978. Principal Exposures of the RiesMeteorite Crater in Southern Germany.München: Bayerisches GeologischesLandesamt.

Chapman, W. 1758. An account of the fossilebones of an allegator, found on the sea-shore, near Whitby in Yorkshire.Philosophical Transactions of the RoyalSociety of London 50, pp. 688–691.

Chen, Z. Q., Shi, G. R. and Kaiho, K. 2002. A

new genus of rhynchonellid brachiopodfrom the Lower Triassic of South China andimplications for timing the recovery ofBrachiopoda after the end-Permian massextinction. Palaeontology 45, pp. 149–164.

Claoue-Long, J. C., Zhang, Z. C., Ma, G. G. andDu, S. H. 1991. The age of the Permian-Triassic boundary. Earth and PlanetaryScience Letters 105, pp. 182–190.

Clark, D. L., Cheng-Yuan, W., Orth, C. S. andGilmore, J. S. 1986. Conodont survival andlow iridium abundances across thePermian-Triassic boundary. Science 233,pp. 984–986.

Clemens, E. S. 1986. Of asteroids anddinosaurs: The role of the press in theshaping of scientific debate. Social Studiesof Science 16, pp. 421–456.

—— 1994. The impact hypotheses and popular

science: conditions and consequences ofinterdisciplinary debate. In Glen, W. (ed.),Mass Extinction Debates: How ScienceWorks in a Crisis, pp. 92–120. Stanford,Ca.: Stanford University Press.

Colbert, E. H. 1968. Men and Dinosaurs. NewYork: Dutton.

Collie, M. and Diemer, J. 1995. Murchison inMoray: a geologist on home ground, withthe correspondence of Roderick ImpeyMurchison and the Rev. Dr. George Gordonof Birnie. Transactions of the AmericanPhilosophical Society 85(3), pp. 1–263.

Courtillot, V. E. 1990. What caused the massextinction? A volcanic eruption. ScientificAmerican 263(10), pp. 85–92.

Cuppy, W. 1964. How to Become Extinct. NewYork: Dover.

Cuvier, G. 1812. Recherches sur les ossemensfossiles de quadrupèdes, où l’on rétablit

les caractères de plusieurs espècesd’animaux que les révolutions du globeparoissent avoir détruites. 4 vols. Paris:G. Dufour et d’Ocagne. 2nd edition: 1821–24; slightly revised 3rd edition, 1825: 4thedition, 1834–36.

—— 1825. Recherches sur les ossemensfossiles de quadrupèdes, où l’on rétablitles caractères de plusieurs animaux dontles révolutions du globe ont détruit lesespèces. 3rd edition. 7 vols. Paris: G.Dufour et d’Ocagne.

Darwin, C. 1859. On the origin of species bymeans of natural selection, or thepreservation of favoured races in thestruggle for life. London: John Murray.

—— 1861. On the origin of species by meansof natural selection, or the preservationof favoured races in the struggle for life.3rd edition. London: John Murray.

Desmond, A. J. 1982. Archetypes andAncestors: Palaeontology in VictorianLondon, 1850–1875. Chicago: Universityof Chicago Press.

Dickens, G. R., Paull, C. K. and Wallace, P.1997. Direct measurement of in situmethane quantities in a large gas-hydratereservoir. Nature 385, pp. 426–428.

Dunbar, C. O. 1940. The type Permian: itsclassification and correlation. Bulletin ofthe American Association of PetroleumGeologists 24, pp. 237–281.

Durant, G. P. and Rolfe, W. D. I. 1984. WilliamHunter (1718–1783) as natural historian:his ‘geological’ interests. Earth SciencesHistory 3, pp. 9–24.

Efremov, I. A. 1937. [On the stratigraphicdivisions of the continental Permian andTriassic of the USSR, based on tetrapodfaunas.] Doklady Akademii Nauk SSSR 16,

pp. 125–132.—— 1940. Kürze Übersicht über die Formen

der Perm- und Trias Tetrapoden-Fauna derUdSSR. Centralblatt für Mineralogie,Geologie und Paläontologie, AbtheilungB, pp. 372–383.

—— 1941. [Short survey of faunas of Permianand Triassic Tetrapoda of the USSR.]Sovetskaya Geologiya 5, pp. 96–103.

—— 1952. [On the stratigraphy of the Permianred beds of the USSR based on terrestrialvertebrates.] Izvestiya AN SSSR, SeriyaGeologicheskaya 6, pp. 49–75.

Ehrlich, P. R. and Ehrlich, A. H. 1990. ThePopulation Explosion. New York: Simon& Schuster.

Erwin, D. H. 1993. The Great PaleozoicCrisis: Life and Death in the Permian.New York: Columbia University Press.

—— 1994. The Permo-Triassic extinction.

Nature 367, pp. 231–236.—— and Pan, H.-Z. 1996. Recoveries and

radiations: gastropods after the Permo-Triassic mass extinction. In Hart, M. B.(ed.) Biotic Recovery from MassExtinction Events, pp. 223–229.Geological Society Special Publication No.102.

Erwin, T. 1982. Tropical forests: their richnessin Coleoptera and other arthropod species.Coleopterists’ Bulletin 36, pp. 74–75.

—— 1983. Beetles and other insects oftropical forest canopies at Manaus, Brazil,sampled by insecticidal fogging. In Sutton,S. L., Whitmore, T. C. and Chadwick, A. C.(eds), Tropical Rain Forest: Ecology andManagement, pp. 59–75. London:Blackwell.

Eshet, Y., Rampino, M. R. and Visscher, H.1995. Fungal event and palynological

record of ecological crisis and recoveryacross the Permian-Triassic boundary.Geology 23, pp. 967–970.

Farley, K.A. and Mukhopadhyay, S. 2001. Anextraterrestrial impact at the Permian-Triassic boundary? Science 293, 2343a(U1-2).

Fitton, W. H. 1839. Elements of geology, byCharles Lyell, Esq., F.R.S. EdinburghReview 69, pp. 406–466.

Forey, P. L., Humphries, C. J., Kitching, I. J.,Scotland, R. W., Siebert, D. J. andWilliams, D. M. 1998. Cladistics: APractical Course in Systematics. 2ndedition, Oxford: Clarendon Press.

Fortey, R. 2000. Trilobite! Eyewitness toEvolution. London: Harper Collins; NewYork: Knopf.

Ganapathy, R. 1980. A major meteorite impacton the earth 65 million years ago: evidence

from the Cretaceous-Tertiary boundaryclay. Science 209, pp. 921–923.

Geikie, A. 1875. Life of Sir RoderickMurchison … based on his journals andletters with notices of his scientificcontemporaries and a sketch of the riseand growth of Palaeozoic geology inBritain. 2 vols. London: John Murray.

Gillispie, C. C. 1951. Genesis and Geology.Cambridge, Mass.: Harvard UniversityPress.

Glen, W. (ed.) 1994. Mass ExtinctionDebates: How Science Works in a Crisis.Stanford, Calif.: Stanford University Press.

Gore, A. 1992. Earth in the Balance. New ed.2000. London: Earthscan.

Gould, S. J. 1987. Time’s Arrow, Time’s Cycle.Cambridge, Mass.: Harvard UniversityPress.

—— and Calloway, C. B. 1980. Clams and

brachiopods – ships that pass in the night.Paleobiology 6, pp. 383–396.

Grayson, D. K. 1984. Nineteenth-centuryexplanations of Pleistocene extinctions: areview and analysis. In Martin, P. S. andKlein, R. G. (eds), Quaternary Extinctions,A Prehistoric Revolution, pp. 5–39.Tucson: University of Arizona Press.

Greene, J. C. 1961. The Death of Adam.Evolution and its Impact on WesternThought. New York: Mentor.

Gruszczynski, M., Halas, S., Hoffman, A. andMalkowksi, K. 1989. A brachiopod calciterecord of the oceanic carbon and oxygenisotope shifts at the Permian/Triassictransition. Nature 337, pp. 64–68.

Hallam, A. 1983. Great GeologicalControversies. Oxford: Oxford UniversityPress.

—— 1987. End-Cretaceous extinction event:

argument for terrestrial causation. Science238, pp. 1237–1242.

—— and Wignall, P. 1997. Mass Extinctionsand their Aftermath. Oxford: OxfordUniversity Press.

Hanlon, M. 2001. The great dying. Daily Mail(23 February), p. 13.

Harland, W. B., Cox, A. V., Llewellyn, P. G.,Pickton, C. A. G., Smith, A. G. and Walters,R. 1982. A Geologic Time Scale.Cambridge: Cambridge University Press.

Hawkes, N. 2001. Crash 250 million years agonearly wiped out life. The Times (23February), p. 13.

Hennig, W. 1966. Phylogenetic Systematics.Bloomington: University of Indiana Press.

Hildebrand, A. R., Penfield, G. T., Kring, D. A.,Pilkington, M., Camargo, Z. A., Jacobsen,S. B. and Boynton, W. V. 1991. Chicxulubcrater: a possible Cretaceous/Tertiary

boundary impact crater on the YucatánPeninsula, Mexico. Geology 19, pp. 867–871.

Hillis, D. M., Moritz, C. and Mable, B. K.1996. Molecular Systematics. 2nd edition.Sunderland, Mass.: Sinauer.

Hoffman, A. 1985. Patterns of familyextinction depend on definitions andgeological timescale. Nature 315, pp. 659–662.

—— 1986. Neutral model of Phanerozoicdiversification: implications formacroevolution. Neues Jahrbuch fürGeologie und Paläontologie,Abhandlungen 172, pp. 219–244.

Holser, W. T., Schönlaub, H.-P., Attrep, M., Jr.,Boeckelmann, K., Klein, P., Magaritz, M.,Pak, E., Schramm, J.-M., Stattgegger, K.and Scmöller, R. 1989. A uniquegeochemical record at the Permian/

Triassic boundary. Nature 337, pp. 39–44.Hsu, K. T. 1980. Terrestrial catastrophe caused

by a cometary impact at the end of theCretaceous. Nature 285, pp. 201–203.

—— 1994. Uniformitarianism vs.catastrophism in the extinction debate. InGlen, W. (ed.), Mass Extinction Debates:How Science Works in a Crisis, pp. 217–229. Stanford, Calif.: Stanford UniversityPress.

Huxley, T. H. 1870. On the classification of theDinosauria, with observations on thedinosaurs of the Trias. Quarterly Journalof the Geological Society of London 26,pp. 32–51.

Isozaki, Y. 2001. An extraterrestrial impact atthe Permian-Triassic boundary? Science293, 2343a (U2).

Jablonski, D. and Raup, D. M. 1995. Selectivityof end-Cretaceous marine bivalve

extinctions. Science 268, pp. 389–391.Jakovlev, N. N. 1922. [Extinction and its causes

as a principal question in biology]. Mysl 2,pp. 1–36.

Jastrow, R. 1983. The dinosaur massacre: adouble-barrelled mystery. Science Digest1983 (September), pp. 151–153.

Jefferson, T. 1799. A memoir on the discoveryof certain bones of a quadruped of theclawed kind in the western parts of Virginia.Transactions of the AmericanPhilosophical Society, 4, pp. 246–259.

Jepsen, G. L. 1964. Riddles of the terriblelizards. American Scientist 52, pp. 227–246.

Jin, Y. G., Wang, Y., Wang, W., Shang, Q. H.,Cao, C. Q. and Erwin, D. H. 2000. Patternof marine mass extinction near thePermian-Triassic boundary in south China.Science 289, pp. 432–436.

Kaiho, K., Kajiwara, Y., Nakano, T., Miura, Y.,Kawahata, H., Tazaki, K., Ueshima, M.,Chen, Z. and Shi, G. R. 2001. End-Permiancatastrophe by a bolide impact: evidence ofa gigantic release of sulfur from the mantle.Geology 29, pp. 815–818.

Karpinskiy, A. P. 1874. GeologischeUntersuchungen im GouvernmentOrenburg. Verhandlungen derKaiserlichen Gesellschaft für dieGesammte Mineralogie 9, pp. 210–212.

—— 1889. Ueber die Ammoneen der Artinsk-Stufe. Mémoires de l’Académie Impérialedes Sciences de St Pétersbourg, 7ème.Série, 37 (2), pp. 1–104.

Kavasch, J. 1986. The Ries Meteorite Crater.A Geological Guide. Donauwörth: Auer.

Keller, G. and Barrera, E. 1990. TheCretaceous/Tertiary boundary impacthypothesis and the paleontological record.

Geological Society of America SpecialPaper 247, pp. 563–575.

Kerr, R. A. 1995. A volcanic crisis for ancientlife. Science 270, pp. 27–28.

King, G. M. 1991a. The aquatic Lystrosaurus:a palaeontological myth. HistoricalBiology 4, pp. 285–321.

—— 1991b. Terrestrial tetrapods and the endPermian mass extinction event. HistoricalBiology 5, pp. 239–255.

—— and Cluver, M. A. 1991. The aquaticLystrosaurus: an alternative lifestyle.Historical Biology, 4, pp. 323–341.

Kitching, J. W. 1977. The distribution of theKaroo vertebrate fauna; with specialreference to certain genera and the bearingof this distribution on the zoning of theBeaufort Beds. Bernard Price Institute forPalaeontological Research, Memoir 1, pp.1–131.

Knoll, A. H. and Bambach, R. K. 2000.Directionality in the history of life:diffusion from the left wall or repeatedscaling of the right? Paleobiology 26(Suppl.), pp. 1–14.

Kummel, B. and Teichert, C. 1966. Relationsbetween the Permian and Triassicformations in the Salt and Trans-Indusranges, West Pakistan. Neues Jahrbuch fürGeologie und Paläontologie,Abhandlungen 125, pp. 297–333.

—— and —— 1973. The Permian-Triassicboundary in Central Tethys. In Logan, A.and Hills, L. V. (eds), The Permian andTriassic Systems and their MutualBoundary. Memoir 2, Canadian Society ofPetroleum Geologists, Calgary, pp. 17–34.

Lai, X. L., Wignall, P. B. and Zhang, K. X.2001. Palaeoecology of the conodontsHindeodus and Clarkina during the

Permian-Triassic transitional period.Palaeogeography, Palaeoclimatology,Palaeoecology 171, pp. 63–72.

Lewin, R. 1983. Extinctions and the history oflife. Science 221, pp. 935–937.

—— 1985a. Catastrophism not yet dead.Science 229, p. 640.

—— 1985b. Catastrophism not yet dead.Science 230, p. 8.

Linnaeus, C. 1753. Species plantarum, 2volumes. Stockholm: L. Salvii.

—— 1758. Systema Naturae per Regna TriaNaturae, Secundum Classes, Ordines,Genera, Species, cum Characteribus,Differentiis, Synonymis, Locis. 10thedition. Stockholm: L. Salvii.

Lomborg, B. 2001. The SkepticalEnvironmentalist. Cambridge: CambridgeUniversity Press.

Loomis, F. B. 1905, Momentum in variation.

American Naturalist 39, pp. 839–843.Looy, C. V., Twitchett, R. J., Dilcher, D. L.,

Van Konijnenburg-Van Cittert, H. A. andVisscher, H. 2001. Life in the end-Permiandead zone. Proceedings of the NationalAcademy of Science 98, pp. 7879–7883.

Lyell, C. 1830–1833. Principles of geology,being an attempt to explain the formerchanges of the Earth’s surface, byreference to causes now in operation. 3vols. London: John Murray.

—— 1838. Elements of Geology. London:John Murray.

McKinney, M. L. 1995. Extinction selectivityamong lower taxa – gradational patterns andrarefaction error in extinction estimates.Paleobiology 21, pp. 300–313.

MacLeod, K. G., Smith, R. M. H., Koch, P. L.and Ward, P. D. 2000. Timing of mammal-like reptile extinctions across the Permian-

Triassic boundary in South Africa. Geology28, pp. 227–230.

McRoberts, C. A. 2001. Triassic bivalves andthe initial marine Mesozoic revolution: arole for predators? Geology 29, pp. 359–362.

Maddox, J. 1985. Periodic extinctionsundermined. Nature 315, p. 627.

Marsh, O. C. 1882. Classification of theDinosauria. American Journal of Science,Series 3 23, pp. 81–86.

—— 1895. On the affinities and classificationof the dinosaurian reptiles. AmericanJournal of Science, Series 3 50, pp. 483–498.

Matthew, W. D. 1921, Fossil vertebrates andthe Cretaceous-Tertiary problem, AmericanJournal of Science, Series 5 2, pp. 209–227.

Maxwell, W. D. 1992. Permian and Early

Triassic extinction of nonmarine tetrapods.Palaeontology 35, pp. 571–583.

—— and Benton, M. J. 1990. Historical testsof the absolute completeness of the fossilrecord of tetrapods. Paleobiology 16, pp.322–335.

May, R. M. 1990. How many species?Philosophical Transactions of the RoyalSociety, Series B 330, pp. 293–304.

—— 1992. How many species inhabit theEarth? Scientific American 267(4), pp. 18–24.

Maynard Smith, J. and Szathmáry, E. 1995. TheMajor Transitions in Evolution. Oxford:W. H. Freeman Spektrum.

Meyer, H. von 1866. Reptilien aus demKupfer-Sandstein des West-UralischenGouvernements Orenburg.Palaeontographica 15, pp. 97–130.

Moodie, R. L., 1923. Paleopathology. Urbana,

Illinois: University of Illinois Press.Morrell, J. B. and Thackray, A. 1981.

Gentlemen of Science. Early Years of theBritish Association for the Advancementof Science. Oxford: Clarendon Press.

Müller, L. 1928. Sind die Dinosaurier durchVulkanausbruche ausgeratet worden?Unsere Welt 20, pp. 144–146.

Mundil, R., Metcalfe, I., Ludwig, K. R., Renne,P. R., Oberli, F. and Nicoll, R. S. 2001.Timing of the Permian-Triassic bioticcrisis: implications from new zircon U/Pbdata (and their limitations). Earth andPlanetary Science Letters 187, pp. 131–145.

Murchison, R. I. 1839. The Silurian System,founded on geological researches in thecounties of Salop, Hereford, Radnor,Montgomery, Caermarthen, Brecon,Pembroke, Monmouth, Gloucester,

Worcester and Stafford; with descriptionsof the coal-fields and overlyingformations. 2 vols. London: John Murray.

—— 1841a. First sketch of some of theprincipal results of a second geologicalsurvey of Russia, in a letter to M. Fischer.Philosophical Magazine and Journal ofScience, Series 3, 19, pp. 417–422.

—— 1841b. Geologicheskaya naogyudeniya vRossii; pis’mo G. Murchisona k’G. Fisherafon Vald’heimu. Gorny Zhurnal, Moskva,1, pp. 160–169.

—— 1842a. Letter to M. Fischer de Waldheim… containing some of the results of hissecond geological survey of Russia.Edinburgh New Philosophical Journal,32, pp. 99–103.

—— 1842b. Anniversary address of thePresident. Proceedings of the GeologicalSociety of London, 3, pp. 637–687.

—— and Verneuil, E. de. 1841a. On thestratified deposits which occupy thenorthern and central regions of Russia.Report of the British Association for theAdvancement of Science, 1840, pp. 105–110.

—— and —— 1841b. On the geologicalstructure of the northern and centralregions of Russia. Proceedings of theGeological Society of London, 3, pp. 398–408.

—— and —— 1842. A second geologicalsurvey of Russia in Europe. Proceedings ofthe Geological Society of London, 3, pp.717–730.

——, —— and Keyserling, A. von 1842. Onthe geological structure of the UralMountains. Proceedings of the GeologicalSociety of London, 3, pp. 742–753.

——, —— and —— 1845. The geology of

Russia in Europe and the UralMountains. 2 vols. Volume 1, London:John Murray. Volume 2, Paris: Bertrand.

Newell, A. J., Tverdokhlebov, V. P. and Benton,M. J. 1999. Interplay of tectonics andclimate on a transverse fluvial system,Upper Permian, southern Uralian forelandbasin, Russia. Sedimentary Geology 127,pp. 11–29.

Newell, N. D. 1952. Periodicity in invertebrateevolution. Journal of Paleontology 26, pp.371–385.

—— 1967. Revolutions in the history of life.Scientific American 208, pp. 76–92.

Nopcsa, F. 1911. Notes on British dinosaurs.Part IV: Stegosaurus priscus, sp. nov.,Geological Magazine (5) 8, pp. 143–153.

—— 1917. Über Dinosaurier, Centralblatt fürMineralogie, Geologie, undPaläontologie 1917, pp. 332–351.

Norell, M. A. and Novacek, M. J. 1992. Thefossil record and evolution: comparingcladistic and paleontologic evidence forvertebrate history. Science 255, pp. 1691–1693.

Novotny, V., Basset, Y., Miles, S. E., Weiblen,G. D., Bremer, B., Cizek, L. and Drozd, P.2002. Low host specificity of herbivorousinsects in a tropical forest. Nature 416, pp.841–844.

Ochev, V. G. and Surkov, M. A. 2000. Thehistory of excavation of Permo-Triassicvertebrates from eastern Europe. In Benton,M. J., Unwin, D. M., Shishkin, M. A. andKurochkin, E. N. (eds), The Age ofDinosaurs in Russia and Mongolia.Cambridge: Cambridge University Press.

Officer, C. B., Hallam, A., Drake, C. L. andDevine, J. D. 1987. Late Cretaceous andparoxysmal Cretaceous-Tertiary

extinctions. Nature 326, pp. 143–149.Oldroyd, D. R. 1990. The Highlands

Controversy. Constructing GeologicalKnowledge through Field-work inNineteenth-Century Britain. Chicago:University of Chicago Press.

Olson, E. C. 1955. Parallelism in the evolutionof the Permian reptilian faunas of the Oldand New Worlds. Fieldiana, Zoology 37,pp. 385–401.

—— 1957. Catalogue of localities of Permianand Triassic vertebrates of the territories ofthe U.S.S.R. Journal of Geology 65, pp.196–226.

—— 1990. The Other Side of the Medal: APaleobiologist Reflects on the Art andSerendipity of Science. Blacksburg, Va.:McDonald & Woodward.

Osborne, R. 1998. The Floating Egg.Episodes in the Making of Geology.

London: Jonathan Cape.Outram, D. 1984. Georges Cuvier. Vocation,

Science and Authority in Post-Revolutionary France. Manchester:Manchester University Press.

Owen, R. 1842. Report on British fossilreptiles. Report of the British Associationfor the Advancement of Science 1841, pp.60–204.

—— 1845a. Description of certain fossilcrania, discovered by A. G. Bain, Esq., insandstone rocks at the south-easternextremity of Africa, referable to differentspecies of an extinct genus of Reptilia(Dicynodon), and indicative of a new tribeor sub-order of Sauria. Quarterly Journalof the Geological Society of London 1, pp.318–322, and Transactions of theGeological Society of London, Series 2 7,pp. 59–84.

—— 1845b. Professor Owen upon certainsaurians of the Permian rocks. Page 637 inMurchison et al. (1845).

—— 1876. Evidences of theriodontselsewhere than in South Africa. QuarterlyJournal of the Geological Society ofLondon 32, pp. 352–363.

Phillips, J. 1838. Geology. Penny Cyclopedia11, pp. 127–151.

—— 1840a. Organic remains. PennyCyclopedia 16, pp. 487–491.

—— 1840b. Palaeozoic series. PennyCyclopedia 17, pp. 153–154.

—— 1841. Figures and descriptions of thePalaeozoic fossils of Cornwall, Devonand west Somerset; observed in thecourse of the Ordnance GeologicalSurvey of that district. London: Longman.

—— 1860. Life on Earth. Its Origin andSuccession. Cambridge: Macmillan.

Pimm, S. L., Russell, G. J., Gittelman, J. L. andBrooks, T. M. 1995. The future ofbiodiversity. Science 269, pp. 347–350.

Pope, A. 1993. Alexander Pope [selections].Edited by P. Rogers. Oxford: OxfordUniversity Press.

Prothero, D. R. 1990. Interpreting theStratigraphic Record. New York: W. H.Freeman.

Rampino, M. R. and Adler, A. C. 1998.Evidence for abrupt latest Permian massextinction of foraminifera: results of testsfor the Signor-Lipps effect. Geology 26,pp. 415–418.

Raup, D. M. 1979. Size of the Permo-Triassicbottleneck and its evolutionaryimplications. Science 206, pp. 217–218.

—— 1986. The Nemesis Affair. New York:Norton.

—— and Jablonski, D. 1993. Geography of

end-Cretaceous bivalve extinctions.Science 260, pp. 971–973.

—— and Sepkoski, J. J., Jr. 1982. Massextinctions in the marine fossil record.Science 215, pp. 1501–1503.

—— and —— 1984. Periodicities ofextinctions in the geologic past.Proceedings of the National Academy ofSciences, U.S.A. 81, pp. 801–805.

Reichow, M. K., Saunders, A. D., White, R. V.,Pringle, M. S., Al’Mukhamedov, A. I.,Medvedev, A. I. and Kirda, N. P. 2002.40Ar/39Ar dates from the West Siberianbasin: Siberian flood basalt provincedoubled. Science 296, pp. 1846–1849.

Renne, P. R. and Basu, A. R. 1991. Rapideruption of the Siberian Traps flood basaltsat the Permo-Triassic boundary. Science253, pp. 176–179.

—— Zhang, Z., Richardson, M. A., Black, M. T.

and Basu, A. R. 1995. Synchrony and causalrelations between Permo-Triassic boundarycrises and Siberian flood volcanism.Science 269, pp. 1413–1416.

Retallack, G. J. 1999. Postapocalypticgreenhouse paleoclimate revealed byearliest Triassic paleosols in the SydneyBasin, Australia. Bulletin of the GeologicalSociety of America 111, pp. 52–70.

——, Seyedolali, A., Krull, E. S., Holser, W.T., Ambers, C. P. and Kyte, F. T. 1998.Search for evidence of impact at thePermian-Triassic boundary in Antarcticaand Australia. Geology 26, pp. 979–982.

——, Veevers, J. J. and Morante, R. 1996.Global coal gap between Permian-Triassicextinction and Middle Triassic recovery ofpeat-forming plants. Bulletin of theGeological Society of America 108, pp.195–207.

Rhodes, F. H. T. 1967. Permo-Triassicextinction. In Harland, W. B. The FossilRecord (ed.), pp. 57–76. GeologicalSociety of London.

Rodland, D. L. and Bottjer, D. J. 2001. Bioticrecovery from the end-Permian massextinction: behavior of the inarticulatebrachiopod Lingula as a disaster taxon.Palaios 16, pp. 95–101.

Rubidge, B. S. (ed.) 1995. Biostratigraphy ofthe Beaufort Group (Karoo Supergroup),South Africa. Pretoria: Council ofGeoscience.

Rudwick, M. J. S. 1969. The strategy of Lyell’sPrinciples of geology. Isis 61, pp. 5–33.

—— 1975. Caricature as a source for thehistory of science: De la Beche’s anti-Lyellian sketches of 1831. Isis 66, pp.534–560.

—— 1976. The Meaning of Fossils. Episodes

in the History of Palaeontology. 2ndedition. New York: Science HistoryPublications.

—— 1985. The Great DevonianControversy: The Shaping of ScientificKnowledge Among GentlemanlySpecialists. Chicago: Chicago UniversityPress.

—— 1997. Georges Cuvier, Fossil Bones,and Geological Catastrophes. Chicago:Chicago University Press.

Rupke, N. A. 1994. Richard Owen, VictorianNaturalist. New Haven: Yale UniversityPress.

Russell, D. A., and Tucker, W. 1971.Supernovae and the extinction of thedinosaurs. Nature 229, pp. 553–554.

Ruzhentsev, V. E. 1950. [Upper Carboniferousammonoids from the Urals.] TrudyPaleontologischeskogo Instituta AN SSSR

29, pp. 1–223.Salvador, A. (ed.) 1994. International

Stratigraphic Guide. A Guide toStratigraphic Classification,Terminology, and Procedure. 2nd edition.International Union of GeologicalSciences.

Schindewolf, O. H. 1950. Grundfragen derPaläontologie. Geologische Zeitmessung– Organische Stammesentwicklung –Biologische Systematik. Stuttgart:Schweizerbart.

—— 1958. Zur Aussprache über die grossenerdgeschichtlichen Faunenschnitte und ihreVerursachung. Neues Jahrbuch fürGeologie und Paläontologie, Monatshefte1958, pp. 270–279.

—— 1963. Neokatastrophismus? Zeitschriftder Deutschen Geologischen Gesellschaft114, pp. 430–445.

—— 1993. Basic Questions in Paleontology.Geologic Time, Organic Evolution, andBiological Systematics. (Judith Shaefertrans.). Chicago: University of ChicagoPress.

Secord, J. A. 1986. Controversy in VictorianGeology: The Cambrian-Silurian Dispute.Princeton: Princeton University Press.

Sedgwick, A. 1831. Address to the GeologicalSociety, delivered on the evening of theanniversary, Feb. 18, 1831. Proceedings ofthe Geological Society of London, 1, pp.281–316.

—— and Murchison, R. I. 1839. Classificationof the older stratified rocks of Devonshireand Cornwall. Philosophical Magazineand Journal of Science, Series 3, 14, pp.241–260.

—— and —— 1840. On the classification anddistribution of the older or Palaeozoic

rocks of the north of Germany and ofBelgium, as compared with formations ofthe same age in the British Isles.Proceedings of the Geological Society ofLondon, 3, pp. 300–311.

Seeley, H. G. 1894. Researches on thestructure, organization, and classificationof the fossil Reptilia. Part VIII. Furtherevidences of Deuterosaurus andRhopalodon from the Permian rocks ofRussia. Philosophical Transactions of theRoyal Society, Series B 185, pp. 663–717.

Sennikov, A. G. 1996. Evolution of thePermian and Triassic tetrapod communitiesof Eastern Europe. Palaeogeography,Palaeoclimatology, Palaeoecology 120,pp. 331–351.

Sepkoski, J. J., Jr. 1984. A kinetic model ofPhanerozoic taxonomic diversity. III. Post-Paleozoic families and mass extinctions.

Paleobiology 10, pp. 246–267.—— 1993. Ten years in the library: how

changes in taxonomic data bases affectperception of macroevolutionary pattern.Paleobiology 19, pp. 43–51.

—— 1996. Patterns of Phanerozoicextinction: a perspective from global databases. In Walliser, O. H. (ed.) GlobalEvents and Event Stratigraphy, pp. 35–52.Berlin: Springer-Verlag.

Sheehan, P. M., Fastovsky, D. E., Hoffman, R.G., Berghaus, C. B. and Gabriel, D. L. 1991.Sudden extinction of the dinosaurs: latestCretaceous upper Great Plains, U.S.A.Science 254, pp. 835–839.

Shoemaker, E. M. and Chao, E. C. T. 1961.New evidence for the impact origin of theRies Basin, Bavaria, Germany. Journal ofGeophysical Research 66, pp. 3371–3378.

Signor, P. W. and Lipps, J. H. 1982. Sampling

bias, gradual extinction patterns andcatastrophes in the fossil record. SpecialPaper of the Geological Society ofAmerica 190, pp. 291–296.

Simpson, S. 2001. Deeper impact. ScientificAmerican 276 (5), pp. 13–14.

Sloan, R. E., Rigby, I. K., Jr., Van Valen, L. M.and Gabriel, D. 1986. Gradual dinosaurextinction and simultaneous ungulateradiation in the Hell Creek Formation.Science 232, pp. 629–632.

Smit, J. and Hertogen, J. 1980. Anextraterrestrial event at the Cretaceous-Tertiary boundary. Nature 285, pp. 198–200.

Smith, F. D. M., May, R. M., Pellew, R.,Johnson, T. H. and Walter, K. R. 1993. Howmuch do we know about the currentextinction rate? Trends in Ecology andEvolution 8, pp. 375–378.

Smith, R. M. H. and Ward, P. D. 2001. Patternof vertebrate extinctions across an eventbed at the Permian-Triassic boundary in theKaroo Basin of South Africa. Geology 29,pp. 1147–1150.

Stafford, R. A. 1989. Scientist of Empire: SirRoderick Murchison, ScientificExploration and Victorian Imperialism.Cambridge: Cambridge University Press.

Stanley, S. M. 1984. Temperature and bioticcrises in the marine realm. Geology 12, pp.205–208.

—— 1986. Earth and Life Through Time.New York: W. H. Freeman.

—— 1988. Paleozoic mass extinctions: sharedpatterns suggest global cooling as acommon cause. American Journal ofScience 288, pp. 334–352.

—— and Yang, X. 1994. A double massextinction at the end of the Paleozoic era.

Science 266, pp. 1340–1344.Stauffer, R. C. (ed.) 1975. Charles Darwin’s

Natural Selection: Being the Second Partof his Big Species Book Written from 1856to 1858. Cambridge: Cambridge UniversityPress.

Stöffler, D. and Ostertag, R. 1983. The Riesimpact crater. Fortschritte derMineralogie 61 (2), pp. 71–116.

Strangways, W. H. T. F. 1822. An outline of thegeology of Russia. Transactions of theGeological Society of London, Series 2 1,pp. 1–39.

Sun, Y., Xu, D., Zhang, Q., Yang, Z., Sheng, J.,Chen, C., Rui, L., Liang, X., Zhao, J. and He,J. 1984. The discovery of an iridiumanomaly in the Permian-Triassic boundaryclay in Changxing, Zhejiang, China and itssignificance. In Developments inGeoscience, Contributions, pp. 235–245.

27th International Geological Congress.Beijing: Science Press.

Surkov, M. V. 2002. Lystrosaurus georgi andthe habits of the lystrosaurs.Palaeontology, in review.

Teichert, C. 1990. The Permian-Triassicboundary revisited. In Kauffman, E. G. andWalliser, O. H. (eds) Extinction Events inEarth History, pp. 199–238. Berlin:Springer Verlag.

—— and Kummel, B. 1976. Permian-Triassicboundary in the Kap Stosch area, EastGreenland. Meddelelser om Grønland597, pp. 1–54.

Terry, K. D. and Tucker, W. H. 1968.Biological effects of supernova. Science159, pp. 421–423.

Thackray, J. C. 1978. R. I. Murchison’sGeology of Russia (1845). Journal of theSociety for the Bibliography of Natural

History 8, pp. 421–433.Tverdokhlebov, V. P. 1971. [On Early Triassic

proluvial deposits of the Pre-Urals, andtimes of folding and mountain-buildingprocesses in the southern Urals.] IzvestiyaAN SSSR, Seriya Geologicheskaya1971(4), pp. 42–50.

——, Tverdokhlebova, G. I., Benton, M. J. andStorrs, G. W. 1997. First record offootprints of terrestrial vertebrates fromthe Upper Permian of the Cis-Urals,Russia. Palaeontology 40, pp. 157–166.

——, ——, Surkov, M. V. and Benton, M. J.2002. Tetrapod localities from the Triassicof the SE of European Russia. EarthScience Reviews 59, in press.

Twelvetrees, W. H. 1880. On a new theriodontreptile (Cliorhizodon orenburgensis,Twelvetr.) from the Upper Permiancupriferous sandstones of Kargala, near

Orenburg in south-eastern Russia.Quarterly Journal of the GeologicalSociety of London 36, pp. 540–543.

—— 1882. On the organic remains from theUpper Permian strata of Kargala, in easternRussia. Quarterly Journal of theGeological Society of London 38, pp.490–501.

Twitchett, R. J. 1999. Palaeoenvironments andfaunal recovery after the end-Permian massextinction. Palaeogeography,Palaeoclimatology, Palaeoecology 154,pp. 27–37.

——, Looy, C. V., Morante, R., Visscher, H.and Wignall, P. B. 2001. Rapid andsynchronous collapse of marine andterrestrial ecosystems during the end-Permian biotic crisis. Geology 29, pp.351–354.

Valentine, J. W. and Moores, E. M. 1973.

Provinciality and diversity across thePermian-Triassic boundary. In Logan, A.and Hills, L. V. (eds)The Permian andTriassic Systems and their MutualBoundary, pp. 759–766. Canadian Societyof Petroleum Geology, Memoir No. 2.

Van Valen, L. M. 1984. Catastrophes,expectations, and the evidence.Paleobiology 10, pp. 121–137.

Wang, X.-D. and Sugiyama, T. 2000. Diversityand extinction patterns of Permian coralfaunas of China. Lethaia 33, pp. 285–294.

Ward, P. D. 1990. The Cretaceous/Tertiaryextinctions in the marine realm: a 1990perspective. Geological Society ofAmerica Special Paper 247, pp. 425–432.

——, Montgomery, D. R. and Smith, R. 2000.Altered river morphology in South Africarelated to the Permian-Triassic extinction.Science 289, pp. 1740–1743.

Watson, D. M. S. 1952. Dr Robert Broom,F.R.S. Obituaries of Fellows of the RoyalSociety 8, pp. 37–70.

—— 1957. The two great breaks in the historyof life. Quarterly Journal of theGeological Society, London 112, pp. 435–444.

Wieland, G. R. 1925. Dinosaur extinction.American Naturalist 59, pp. 557–565.

Wignall, P. B. 2001. Large igneous provincesand mass extinctions. Earth-ScienceReviews 53, pp. 1–33.

—— and Hallam, A. 1992. Anoxia as a cause ofthe Permian/Triassic extinction: facies asevidence from northern Italy and thewestern United States. Palaeogeography,Palaeoclimatology, Palaeoecology 93, pp.21–46.

—— and —— 1993. Griesbachian (earliestTriassic) palaeoenvironmental changes in

the Salt Range, Pakistan and southwestChina and their bearing on the Permo-Triassic mass extinction.Palaeogeography, Palaeoclimatology,Palaeoecology 102, pp. 215–237.

——, Kozur, H., and Hallam, A. 1996. Thetiming of palaeoenvironmental changes atthe Permo-Triassic (P/Tr) boundary usingconodont biostratigraphy. HistoricalBiology 12, pp. 39–62.

—— and Twitchett, R. J. 1996. Oceanic anoxiaand the end Permian mass extinction.Science 272, pp. 1155–1158.

——, —— 2002 Extent, duration and nature ofthe Permian-Triassic superanoxic event.Geological Society of America SpecialPaper 356, pp. 395–413.

Williams, J. S. 1938. Pre-congress Permianconference in the U.S.S.R. Bulletin of theAmerican Association of Petroleum

Geologists 22, pp. 771–776.Wilson, E. O. 1992. The Diversity of Life.

Cambridge, Mass.: Harvard UniversityPress; London: Penguin.

—— 2002. The Future of Life. New York,Alfred A. Knopf; London: Little Brown.

—— and Peter, F. M. (eds) 1988. Biodiversity.Washington, D.C.: National AcademyPress.

Winchester, S. 2001. The Map that Changedthe World. The Tale of William Smith andthe Birth of a Science. London: PenguinViking; New York: HarperCollins.

Woodward, A. S. 1898, Outlines of VertebratePaleontology for Students of Geology.Cambridge: Cambridge University Press.

—— 1910, Presidential Address to Section C,Report of the British Association for theAdvancement of Science, 1909, pp. 462–471.

Wooler, -. 1758. A description of the fossilskeleton of an animal found in the alumrock near Whitby. PhilosophicalTransactions of the Royal Society ofLondon, 50, pp. 786–791.

Xu, D., Na, L., Chi, Z., Mao, X., Su, Y., Zhang,Q. and Yong, Z. 1985. Abundance ofiridium and trace metals at thePermian/Triassic boundary at Shangsi inChina. Nature 314, pp. 154–156.

Yang, Z., Sheng, J. and Yin, H. 1995. ThePermian-Triassic boundary: the globalstratotype section and point. Episodes 18,pp. 49–53.

Yin, H., Huang, S., Zhang, K., Hansen, H. J.,Yang, F., Ding, M. and Bie, X. 1992. Theeffects of volcanism on the Permo-Triassicmass extinction in South China. In Sweet,W. C., Yang, Z., Dickins, J. M. and Yin, H.(eds) Permo-Triassic Events in Eastern

Tethys, pp. 146–157. Cambridge:Cambridge University Press.

Zhou, L. and Kyte, F. 1988. The Permian-Triassic boundary event: a geochemicalstudy of three Chinese sections. Earth andPlanetary Science Letters 90, pp. 411–421.

Zittel, K. A. von. 1901. History of Geologyand Palaeontology to the End of theNineteenth Century. London: Walter Scott.

ILLUSTRATION CREDITS

All chapter heading details are taken from thedrawings in the book by John Sibbick

1 Drawing by John Sibbick2 Drawing by John Sibbick3 Drawing by John Sibbick4 From Murchison 1841a5 Drawing by John Sibbick, from various

sources6 From Phillips 18607 Drawing by John Sibbick8 ‘Professor Ichthyosaurus’ by Henry De La

Beche, 18309 Drawing by John Sibbick10 Drawing by John Sibbick11 Based on Alvarez et al., 1980 and other

sources12 Drawing by John Sibbick13 Based on various sources14 Based on various sources15 Based on various sources16 Much modified from Benton, 199517 Based on data in Benton, 1995; redrawn

and embellished by John Sibbick18 Modified from Raup and Sepkoski, 198219 Modified from Raup and Sepkoski, 198420 Based on data in Keller and Barrera, 199021 Based on data in Maxwell and Benton

199022 Modified from Benton et al., 200023 Based on various sources24 Based on the work of John Scotese, and

other authors25 Based on Wignall and Hallam, 1993

26 Courtesy of Tony Hallam27 Based on Jin et al., 200028 Drawing by John Sibbick29 Drawing by John. Sibbick30 Modified from Gould and Calloway, 198031 Drawing by John Sibbick, based on various

sources32 Drawing by John Sibbick, based on various

sources33 Based on Ochev and Surkov, 200134 M. J. Benton35 Drawing by John Sibbick, based on various

sources36 Drawing by John Sibbick, based on

Bystrov, 195737 Drawing by John Sibbick38 Drawing by John Sibbick39 Based on work during the 1995

expedition, from Newell et al., 199940 Based on the work of Valentin

Tverdokhlebov41 Courtesy of Paul Wignall42 Drawing by John Sibbick43 M. J. Benton44 Based on data in Smith et al., 199345 Courtesy of Tony Hallam and Paul

Wignall

INDEX

All page numbers refer to the 2008 printedition

Page numbers in italics refer to illustrations

Abbeville, France 79acid rain 9, 276, 277, 280, 281acromegaly 84actualism 64adenosine triphosphate 128Adler, Andre 185aestivation 217Agassiz, Louis 77, 78Aidaralash Creek, Aktöbe 162Alberti, Friedrich von 46, 52algae 130, 164, 173, 188, 287, 295

Allosaurus 293alluvial fans 221, 241, 246; see also megafansAlps 61, 77Alvarez, Luis W. 96, 97–98, 99, 100, 101, 110,

112–13, 114–15, 116, 118, 121, 138, 139,257

Alvarez, Walter 96, 97–98Amalitskii, V. P. 214ammonites 25, 44, 45, 73, 75, 89, 91, 133,

134, 161; extinction 147ammonoids 161, 169, 181, 186, 192, 256,

257, 294, 295, 297amphibians 19, 20, 34, 92, 133, 159, 204, 211,

214, 233, 237, 248, 249, 256, 298;diversification of 256; evolution of 13, 14,19; fossils 19, 92, 162, 214; groups 133;species loss during end-Permian 220

anapsids 218Annatherapsidus 233

angiosperms 87anhydrite 265anoxia 88, 133, 134, 163, 166, 176, 177, 179,

193, 201, 250, 269, 275, 280, 281, 294,304; see also low oxygen conditions andsuperanoxia

anoxic conditions 168–69, 175, 293, 294;waters 296

Antarctica 205, 206anthracosaurs 239anti-Darwinian viewpoints 80, 91, 92, 94; see

also Darwinism and neo-DarwinianismArchaeopteryx 302Archibald, David 144Archosauria 233archosaurs 24, 236, 240–41, 291–92; basal

293Archosaurus 233Arctic Ocean 183, 214

argon 259; argon-argon dating method 170,264

Armenia 163, 165Arthropoda 198–99arthropods 186, 192, 270, 281, 288Artinskian 158, 170Asaro, Frank 96ash 99, 109, 117, 119, 121, 169, 277; bands

170Asselian stage 158, 160asteroid impact see meteorite impactasymptote 287Atlantic Ocean 176, 254Audova, Alexander 85Audubon, John James 289Aves 142Azerbaijan 163, 165

Bain, Andrew Geddes 206–8, 212, 218Bain, Thomas 207–8Bakker, Robert 113Baksi, Ajoy 264Baoqing Member 166, 167basalt 262–64, 265, 274; eruptions 121, 271,

277; flood 263, 265, 282, 303; lava 213,262, 281

Batrachia 20Beaufort Group 213, 216, 220, 256; see also

Karoo Basin sequenceBeche, Henry De la 67–68, 67, 206Becker, Luann 11, 259, 260, 261Becquerel, Henri 73beetles 216, 280, 288, 301belemnites 89, 134Bellerophon Formation 162bellerophontid gastropods 164

Benthosuchus 214, 234, 236bentonite 107Berner, Robert 283‘big five’ mass extinctions 124–25, 130–38,

154, 284, 298; Late Devonian 124, 132–33; Late Ordovician 124, 132, 136; LateTriassic 124, 133, 153; see also end-Permian mass extinction, extinctions, KTmass extinction and mass extinctions

biodiversity 9–10, 15, 123, 125–26, 285–89,299, 302, 304

bipedalism 24birds 24, 74, 89, 130, 298, 301; basal 150;

diversity of 125; extinction of 289–90;fossils 109,

bivalves 44, 133, 134, 152, 173, 190, 196–97,197, 256, 270, 294, 295; clams 134;mussels 134; oysters 134; 294, 294;recovery after end-Permian extinction 197,197; see also Claraia, Eumorphotis,

Promyalina and Unionitesblastoids 189, 256Boreal Ocean 177, 181, 293Boreapricea 234Boucher de Perthes, Jacques 79, 80Bowring, Sam 170–71brachiopods 44, 52, 92, 132, 133, 134, 164,

179, 190–91, 191–92, 196–97, 197, 201,202, 281; articulate diversification of 132,133, 256; fossils of 25, 166, 168, 173,177; see also Crurithyrus and Lingula

breccia 105, 107, 119, 120; bunte 106, 109;crystalline 106

Bristol, University of 227, 230Brithopus 33, 34British Association for the Advancement of

Science (BA) 31–32, 33, 47, 82British Museum, London 208, 211Broom, Robert 208–9, 212, 226, 227

Browne, Malcolm 113bryozoans 17, 92, 133, 164, 173, 178, 179,

188, 256, 295Buckland, William 29, 30, 61, 62, 68, 76–77,

101buckyballs 259Buntsandstein 46

California, University of, Berkeley 96, 97, 98Calloway, Jack 196, 197–98Cambrian 48, 49, 50, 69, 128, 129, 131, 132,

151; extinction event in Late 132; radiation180

Camp, Charles L. 93Campbell, Ian 265, 266Canidae 142Capitan Limestone Formation 186, 188Capitanian Stage extinction 256, 267Capitosaurus 215

carbon 176, 259, 269, 272, 274, 277, 280,296; cycling 271, 282, 300; dioxide 87,121, 176, 264, 266, 269, 271, 272, 273,275, 276, 277, 279, 280, 281, 300;isotopes 172, 270, 272, 273–74, 276, 279,283

Carboniferous 20, 24, 49, 51, 52, 53, 73, 129,131, 158, 161, 213; Late 247; rocks 45,48, 52, 159

Carnian stage 292Carroll, Bob 71cartilage 84, 86catastrophic models of extinction 7, 11, 79,

80, 147, 156, 271catastrophism 8–9, 13, 57, 64, 69, 78, 91, 92,

93–95, 101, 103, 109, 113, 114, 157;debate with uniformitarianism 58–62

catastrophists 8–9, 57, 58, 61, 68, 69, 79, 103,111, 113, 143, 154, 252

Cenomanian-Turonian extinction event 134Cenozoic era 47–48, 48, 129, 131centipedes 20, 216, 280, 288cephalopods 132, 133, 134, 193, 294Cetiosaurus 30Chamberlin, T. C. 103Changhsingian Stage 158, 256Chao, E. C. T. 105, 106, 109Chapman, William 26Chasmarosuchus 234, 236‘chert gap’ 184, 186cherts 176, 184, 246; radiolarian 184Chicxulub crater, Mexico 117–21, 118, 120,

303Chhidru Formation 164China 14, 163, 165–66; see also Meishan

sectionChinglung formation 168

chiniquodonts 291, 292chlorine 276, 277Chroniosuchus 233, 239clades 142, 150, 154cladistics 148–49Claeys, Philippe 258clams 134, 190; fossils of 25Claraia 164, 193, 201, 270, 294, 294, 296,

297–98; survival of end-Permian massextinction 181

clay bands 99, 107, 168, 170, 175Clemens, Bill 110Clemens, Elisabeth 113–16climate 14, 78, 84, 132, 133; changes 84, 85,

92, 132, 134, 135, 266, 304; decreasedtemperatures 77, 85, 87, 88, 90, 111;deterioration 71, 121, 185; increasedaridity 87, 88, 90, 241, 245; increasedrainfall 87, 88, 90, 241; increased

temperatures 77, 87, 88, 90, 250club-mosses 224, 279, 280Cluver, Michael 226coal 42–43, 45, 52, 297; forests 20‘coal gap’ 248, 297Coal Measures 42–43, 53cockroaches 216, 300coesite 105collector curve 286, 287comet 88, 138, 257comparative anatomy 24, 27–28, 35competitive replacement 197, 197conglomerates 241, 245–46conifers 130, 216, 222, 224, 268, 292conodont 133, 161, 164, 166, 181; fossils

166, 173; Hindeodus parvus 166, 168,169; Streptognathodus isolatus 162

continental: drift 8, 65, 132, 187, 255; shelf

204, 205; uplift 85Cope, Edward Drinker 76Copper Sandstones, Russia 28, 33, 35, 211,

212, 229; reptile fossils found in 28–29,30, 33, 34, 36, 211

coprolites 76coral 133, 179, 181, 186, 187, 191, 256, 270,

295; Caninia 187; diversification of 132;fossils 25, 177, 178; Heliolites 187; reefs123, 195–96; limestones 159; rugose 173,187, 256; scleractinian 188, 295; skeletons186, 295; tabulate 187; see also reefs

cosmic: dust 97; radiation 88, 92–93, 114; rays71, 93, 94, 255

Courtillot, Vincent 263, 264creationists 7, 15, 194Cretaceous 8, 30, 33, 35, 81, 82, 83, 84, 85,

90, 110, 129, 131, 134, 151, 298; bivalveand gastropod species 152–53; fossils 44,

76, 89, 151; Late 82, 84, 85, 89, 110, 111,154; rocks 31, 51, 74, 89, 144, 151

Cretaceous-Tertiary mass extinction see KTevent

crinoids 92, 133, 164, 177, 188–89crocodiles 13, 18, 24, 26, 32, 95, 111, 152,

233; ancestors of 21, 26, 31, 35crocodilians 133, 291–92Crurithyris 192–93, 201Cuppy, Will 87Cuvier, Georges 27, 29, 30, 72, 94, 101, 123–

24, 305; use of comparative anatomy 27–28, 35; and catastrophism 58–62, 63, 65–66, 79

Cycads 224cynodonts 210, 218, 220, 225, 231, 233, 236,

248, 291Cynosaurus 218

Darwin, Charles 19, 68, 75, 80, 82, 90, 92, 93,94, 127, 195, 196; see also neo-Darwinismand anti-Darwinian viewpoints

dating methods 165, 249; argon-39/argon-40170; for rocks 13, 73, 85, 159; precision of142–43; problems of 72, 75, 116, 130,144, 156; radiometric 73, 121; uranium-lead 170

Deccan Traps, India 121, 263, 264Deuterosaurus 34‘devil’s toenails’ see GryphaeaDevonian 20, 69, 38, 41, 45, 48, 49, 50, 51,

73, 129, 131; Late 132–33, 136diademodonts 291diapsids 24, 218, 248; basal 220, 225Dickens, Gerry 273Dicynodon 207, 209, 216, 218, 225, 233, 277,

278, 279, 280, 281, 292Dicynodon Zone 216–19, 217, 220–21, 233,

236, 277dicynodonts 21–22, 207, 210, 219, 220, 226,

231, 233, 236, 239, 248, 281, 291; basal256; see also Dicynodon, Lystrosaurus,Myosaurus

diluvialist 61, 62dinocephalians 210, 215, 256Dinornithidae 301Dinosauria 31, 32–33dinosaurs 23, 25, 29, 33, 34, 66, 75, 76, 80–

81, 92, 130, 133, 134, 233, 292;appearance of 23, 24; of the Cretaceous 82,83, 85; digestive systems of 86; diseases86; diversity of 76; eggs 85, 86, 87;extinction of 8, 9, 11, 12, 13, 24, 75, 76,80–81, 82, 83–84, 85, 88, 89, 90, 95, 96–122, 144, 152, 184; first identification of29, 30, 33; fossils of 13, 90, 128, 143,147, 213; replacement by mammals 84;theories for the extinction of 84, 85–88,

90, 110; see also individual types and KTmass extintion

disaster species 303–4diversification of life 47, 48, 48, 123, 127,

135, 138, 146, 146, 148, 189, 225, 282,284; in the sea 131, 132, 256; on land 131,225–26; see also biodiversity

DNA 86, 94, 127–28, 149–50Dobzhansky, Theodosius 82dodos 289, 300dragonflies 20, 216dropstones 268Dunbar, Carl O. 160, 161dust clouds 9, 88, 99–100, 101, 108, 116, 265Dvinia 233Dvinosaurus 214dwarfing 201Dwyka Tillite 213

Ecca Group 213echinoderms 132, 133, 188: blastoids 189; sea

urchins 132, 166; sea lilies 132, 133;starfish 132, 166

echinoids 164, 177ecological crisis, current 15–16, 125, 284,

289–90, 300–2ecosystems: evolution of new 130; Permian

20–21, 23; recovery phase of 10, 153, 249,299

Ediacara Hills, Australia 132Efremov, Ivan Antonovich 213–15, 214, 228,

230, 234, 239, 248Ehrlich, Anne 284Ehrlich, Paul 284Eichwald, E. I. von 34, 50ejecta 104, 106, 107, 108, 109El Kef, Tunisia 142elephants 27, 28, 34, 77, 152, 250

endemism 254–55end-Permian mass extinction 4, 19, 21–22, 24,

55, 75, 76, 88, 92, 93, 121, 124, 125, 133,140, 141, 156–57, 162, 168, 175, 181,212, 236, 256, 264, 304; affecting life onland 14, 180, 181, 202–3, 215, 219–28,233, 248, 252, 255, 262, 283, 290, 302;affecting life in the sea 14, 133, 180, 181–203, 219, 294; and extinction of trilobites198–99; carbon isotope shift 270–74;catastrophic theories of causes 7–8, 71, 72,257; causes of 253, 254–55, 257; climaticchanges associated with 14, 271; denial of93, 204, 252; effect on bivalves 196, 197,197; effect on brachiopods 190, 191, 196,197, 197; effect on reefs 186–88; evidencefor in Greenland successions 177; evidencefor in Japanese sections 177–78; fullerenesand 259–61, 282; fungal spike 223; gashydrates and 272–74, 277, 281, 282; globalwarming at 266–68, 273, 276; greenhouse

effect 276–77, 304; increase insedimentation 221–22, 250; iridiumanomaly 257–58; loss of families 140,141, 180, 202, 220, 256; loss of plant life222–24, 250; recognition of 47, 71, 80;recovery period 10, 153, 155, 181, 241,284–85, 290–98, 294, 302, 303;Schindewolf and 92–93, 94, 164, 252; sea-level changes and 254–55; shocked quartzand 258; Siberian Traps and 262–77, 281,282, 304; species loss during 140, 141,180, 202–3, 220, 256, 282; superanoxia269–70; survival after 9–10, 22, 23, 23,181, 189, 201–2, 225–27, 249–50, 294;theory of meteor impact 7–8, 10–11, 12,14, 71, 257–62, 282; three pulses of 256–57; timing of 156–57, 164, 173–76, 177,178–79, 232–33, 255, 256–57, 303; seealso ‘big five’ mass extinctions,extinctions, mass extinctions and Permian

entropy 88

environment, human impact on 8, 15–16, 284,289–90, 300–2

Eocene 154, 298Eocene-Oligocene extinction event 134epidemics 90Erman, Georg Adolf 39, 51Erwin, Doug 124, 156, 165, 173, 180, 253,

260, 264–65Erwin, Terry 288Eshet, Yoram 222–23Eumorphotis 193, 201, 294, 294, 297–98evaporites 46evolution 14, 22, 48, 80, 91, 92, 93, 123, 138,

149, 295, 296, 297; neo-Darwinian modelof 82; of biological communities 114;theories of 194–95

evolutionary trees 148exoskeleton 132

extinctions 7–16, 17, 22, 27, 28, 35, 55, 57,58, 66, 80, 131, 136, 142, 152, 255, 265;attributed to Flood 26, 61, 68, 77, 79;catastrophist view of 8, 35, 58, 70;Christian view of 26; Darwin’s view of 75;denial of 7, 26, 28, 58, 71, 72; end-Capitanian 267; end-Carnian 292; end-Jurassic 134, 190, 296; evidence of infossil record 8, 14, 75, 148; global scale78; in the sea 14, 136, 255; Late Cambrian132; Late Precambrian 132; local scale 75;Lyell’s view of 58, 66, 69; mid-Miocene134; mid-Oligocene 134; of the dinosaurs8, 9–10, 11, 12, 13, 24, 47, 74, 75, 76, 80–81, 82, 83–84, 85, 88, 89, 90, 91, 95, 96–122, 123; patterns of 138, 142–43, 152;Pleistocene 76, 77–78, 79–80, 134–35;rates 135, 136, 156, 282, 289, 290, 302;resistance to 180; role of humans in 8, 78,79, 80, 289; survival of 9–10, 227; see also‘big five’ mass extinctions, end-Permian

mass extinction, KT mass extinction andmass extinctions

extraterrestrial models of extinction 114, 133,134, 256

families 125, 151, 191, 202Farley, K. A. 260, 261Farrar, Edward 264feedback systems 12, 275; negative 271–72,

276, 280; positive 272fern spike 117ferns 57, 87, 117, 216, 222, 224, 279, 280finalism 80, 81fish 19, 66, 129, 132, 133, 134, 150, 184, 200,

201; bony 201; fossils of 25, 26, 61, 105,109, 128, 173, 199–200

Fitton, W. H. 63flies 29, 280Flood, Great 26, 61, 68, 77, 79

fluorine 276, 277food chains 202, 250food webs 201–2, 219‘fool’s gold’ see pyritesforaminifera 89, 134, 142, 164, 169, 179, 181,

184–85, 186, 201; fossils 166, 168, 173,177; survival of end-Permian massextinction 185–86

Forbes, J. D. 69forests 20, 87, 123, 129, 130; destruction of

222, 299, 300fossil fuels 271, 272fossil record 7, 8, 89, 142, 145–46, 146, 147–

48, 154, 155; quality of 75, 93, 124–25,126, 128, 141, 150, 151, 152, 147–48,185, 200, 219–20

fossil sampling 143–44, 173fossils: as basis of stratigraphy 47, 85;

classification of 42; collecting 24, 25, 29,

30, 34, 146, 146, 147, 187, 208–9, 218;early ideas about 25–28

foxes 79Fox-Strangways, W. T. H. 50frogs 19, 34, 111, 226, 234, 301fullerenes 259, 262, 282, 259fungal spike 223, 224, 225, 248, 276fungi 150, 223, 289fusulines 185–86, 256fusulinid fossils 173

gap excess ration (GER) 151Garjainia 240–41gas hydrates 272–73, 274, 275, 279, 281, 282,

304; reservoirs 277gastropods 132, 133, 134, 152, 164, 173, 192,

193, 201, 270genetic code 150genetic mutations 93

genetics 82, 92Geological Society of America 258Geological Society of London 36, 37, 38–39,

41–42, 50, 51, 53, 60, 63, 206, 207geological time-scale 13, 37, 41, 42, 43, 46,

47, 48, 49, 51–52, 69, 72, 73gigantism 87Gilbert, G. K. 103glaciation 77, 132, 213, 267, 276glaciers 66, 77, 213, 267, 268glandular malfunction 90glass: beads 116, 117, 119, 121; bombs 107,

109, 119; natural 107global stratotype 161, 164, 166global warming 250, 267, 271, 275Glossopteris 216, 218, 220, 222, 268, 270gneiss 107goethite 269

‘golden spike’ 157, 160–62, 166, 169Gondwana 205, 213, 222, 254, 267, 268, 292goniatites 169, 173Gore, Al 284, 290, 301, 302gorgonopsian 4, 210–11, 214, 219, 219, 220,

226, 233, 248, 281, 291, 292Gould, Stephen Jay 196, 197–98Gower, David 242gradualism 64, 65, 69, 80, 94, 143, 147, 157gradualist ecological succession model 110–

11granite 24–25, 106, 107, 262, 241great auk 289, 300greenhouse effect 15, 266, 276; runaway 281greenhouse gases 266Greenland 14, 163, 177–78: end-Permian

event in 178–79, 181, 183; early Triassicfauna of 181

Gryphaea 25Guadalupe Mountains, Texas 186, 188Guadalupian reefs, Texas 267Gubbio, Italy 97–98; iridium spike at 98, 99Guryul Ravine, Kashmir 163–64gymnosperms 87gypsum 52, 168, 169, 177

haematite 176, 269haemoglobin 149Hallam, Tony 164, 166, 169, 175, 178, 201,

266, 267, 271Hawkes, Nigel 259helium 259, 260, 262, 282Hell Creek Formation 144Hennig, Willi 148–49Hildebrand, Alan 118–19Hindeodus parvus 166, 168, 169

hippopotamus 21, 76, 77, 79Hitchin, Rebecca 150Hoffman, Toni 202Homo sapiens 15, 127, 142hormone systems 86, 87horsetails 216, 222Hsü, Ken 103Human Genome Project 150humans: ancient remains of 68; as cause of

extinctions 8, 78, 79, 80, 289; impact onenvironment 300–1, 302, 304; migration of79, 80; origins of in Cenozoic 129;population growth 15, 301–2; role inPleistocene mammalian extinction 78, 79,80; stone age 78, 79

Hungarian Geological Survey 84Hunter, William 27Hunterian Museum, London 30Hushan section, China 175

Huxley, Julian 82Huxley, Thomas Henry 75–76, 208hyaenas 76hydrochloric acid 276, 279hydrogen 272; sulphide 272, 281hydrozoans 92Hylaeosaurus 30, 32, 33hypophysis 84Hyracotherium 298–99

ice age 78, 124, 218–19ice sheets 76, 77, 78, 134Iceland, basalt flood eruption 263ichthyosaurs 31, 68, 74, 134Iguanodon 30, 32–33, 57Inostrancevia 214, 233, 233insects 20, 132, 133, 286–87, 287, 288;

diversification of 130;

International Geological CorrelationProgramme 160–61

International Union of Geological Sciences160–61

Iran 163, 165iridium 97, 100, 116, 119, 121, 140, 257, 258,

282; increase in at the KT boundary 98–99,98, 114

iron 163, 239; oxide 269iron pyrites see pyriteisotopes: carbon 172, 270, 272, 273–74, 276,

279, 283; oxygen 171, 267, 273Isozaki, Yukio 260

Jablonski, David 152–53Jastrow, Robert 112jellyfish 132, 145Jepsen, J. L. 86Jin, Y. G. 173, 174–75

Journal of Geophysical Research 105Jurassic 26, 33, 35, 44, 45, 51, 129, 131, 213;

Early 73; extinctions 134, 140; fossils of44, 68, 76; Late 73, 82; Middle 44; rocks25, 29, 30, 31, 44; sediments 108

Kaiho, Kunio 262Kainozoic see Cenozoickangaroos, giant 79Kankrin, Count I. F. 54Karoo Basin, South Africa 205–6, 207, 208,

210, 212, 213, 215–16, 217, 219, 220,227, 228, 230, 236, 256, 277, 302;Lystrosaurus fossils in 225; Permo-Triassic boundary 206; fossils found in206–8, 231, 233, 249; sedimentationpatterns, change in 221, 222

Karpinskiy, A. 157–58Kashmir 163, 165Kathwai Member 164

Kazakhstan 239Kazan University 211Kazanian stage 157Kerr, Richard 258Kerry slug 301Keyersling, Count A. 51, 52King, Gillian 26Kirkdale Cavern, Yorkshire 76Kitching, James 225, 227Kotlassia 214, 233Krakatoa eruption 99–100, 108Kranz, Major W. 104KT boundary 98, 99, 110, 111, 114, 117, 119,

142, 144; clays 99, 100, 101, 116, 117,119; meteorite impact at 144, 157;sections 258; use of isotope analysis at 171

KT mass extinction 89–90, 96–122, 124, 134,140, 152, 153–54, 156, 180, 256, 257,

258, 263, 265, 282, 298, 302; extinction ofbivalves 153; killing off of ammonoids147, 192; recovery period 302; survivors of90, 298; see also ‘big five’ massextinctions and dinosaurs

Kummel Bernhard 163–64, 165, 177; ‘mudballtheory’ 177–78

Kungurian stage 157Kupferschiefer (Zechstein) 34, 43Kutorga, S. S. 33–34, 211

Laki fissure 263Laubenfels, M. W. de 95, 114Laurasia 177, 254lava 7, 11, 12, 63, 104, 117, 119, 121, 206,

213, 262, 263, 266, 277, 281, 303; basalt121, 213, 262, 263, 264, 265, 274, 277,281, 282

Lazarus taxa 200

limestone 46, 99, 109, 117, 119, 165, 166,167, 175, 204, 241, 245, 246, 269;Cretaceous 119; dolomites 162; Eocene59; formation in the tropical belt 267;geochemical composition of 121; Jurassic106; Lias 26; Magnesian 52; marine 45, 46,61, 147, 159; microgastropod grainstone201; Miocene 59; Permian 52

Lingula 192–93, 201, 293, 294Linnaeus, Carolus (Linné, Carl von) 126–27Lipps, Jere 174lizards 13, 19, 21, 24, 29, 32, 34, 111, 133,

209Lomborg, Bjorn 300Longueuil, Baron Charles de 27Loomis, Frederick 82Looy, Cindy 178, 223, 224lophophore 190low oxygen conditions 168, 201, 270, 293,

294, 296; see also anoxia and superanoxiaLower Vetluga Community 248Ludwigia murchisonae 44Luehea seeannii 288lungfishes 216–17Lycaenops 4lycopsids 224Lyell, Charles 56, 57–58, 61–66, 68, 69, 70,

72, 75, 78, 79, 90, 94, 101, 113, 121, 123,124, 195, 304; use of the term ‘uniformity’64–66; Principles of Geology 57–58, 63,64, 68, 69, 206

Lystrosaurus 21–23, 23, 220, 224–27, 226,234, 236, 248, 291, 296

Lystrosaurus Zone 220–21, 225, 236

McKinney, Mike 202magma 25, 253, 265magnetite 262

mammals 24, 28, 60, 61, 74, 76, 77, 78, 84,85, 87, 89, 90, 110, 111, 130, 133, 134,149, 154, 209; bones of 78; diversity of125; early 34, 110; fossil 59, 66, 34, 35;present rate of extinction 289

mammoths 61, 76, 78, 79, 135, 196, 219;bones 25, 76; extinction debate 28, 76;Siberian 27

Mantell, Gideon 30, 33Mantell, Mary 30Maoris 79, 301Marie Celeste 272–73Marsh, Othniel Charles 76marsupials 79, 152, 208mass extinctions 7–16, 21, 22, 23, 47, 55, 71–

72, 74, 75, 76, 80, 81, 96, 100, 121, 122,123, 140, 142, 146, 152, 154–55, 172,200, 249, 256, 257, 263, 271, 299;astronomical causes for 139–40, 139, 264;

common factors to 125, 135–36, 154–55,303; Darwin and 75; debate about 57, 58,69–70, 71–72, 88; definition of 124–25,130, 135–38, 141–42, 154–55; end-Capitanian event 256, 267; flood basalts ascause 263–64; global cooling as cause 266;Late Devonian 124, 132–33; LateOrdovician 124, 132, 136; Late Triassic124, 133, 153; modern studies of 96;periodicity of 138–40; recovery after 153–55, 284–85, 290, 292, 293–99, 300, 303–4; regression analysis 136–38; scale of140–41; Schindewolf and 92–95;selectivity during 152, 154, 181, 201; sixth284, 290; see also ‘big five’ massextinctions, end-Permian mass extinction,extinctions and KT mass extinction

mastodons 27, 76, 78, 135, 196, 219;extinction debate 28; the Ohio incognitum27

Mastodonsaurus 215

Matthew, William Diller 84Maxwell, Des 145–46, 147mayflies 216Mayr, Ernst 82Mazzin Member 163megafans 246–47, 246, 248Megalosaurus 29, 32, 33, 76Megatherium 27; see also slothmeiofauna 287–88Meishan, China 157, 269Meishan Member 167Meishan section, China 163, 165, 166, 168,

173, 174, 175, 176, 178, 181, 182, 255,256, 259, 260, 261, 270; Chinglungformation 168; early Triassic fauna of 181;marine decline during 224; Permo-Triassicboundary 166–69, 167, 172–73, 174, 178;radiolarian cherts in 184; selection asglobal stratotype 166

Mesozoic 47–48, 48, 74, 81, 73, 76, 85. 133,134, 147, 252

metabolic rates 74Meteor Crater, Arizona 103, 104, 105meteorite impact 7–8, 10, 11–12, 14, 22, 58,

61, 64, 65, 71, 88, 99, 101, 103, 106, 112,113–14, 120, 121, 123, 133, 134, 138,139, 140, 257, 262, 303, 304; andextinction of the dinosaurs 8, 9, 13, 95,102, 117; policies relating to prevention of12; pressure during impact 108, 117;responsible for KT event 96–97, 117–21,120, 282, 303; see also Chicxulub Crater,KT mass extinction and Ries Crater,Nördlingen

meteorite shower 97methane 86, 272, 277, 279; burp 273, 282–83;

hydrate 272, 273; poisoning 87; release283

Meyer, Hermann von 30, 211

Meyerdorf, Baron A. von 51mice 79Michel, Helen V. 96microbes 10, 14, 128, 135, 145, 269, 289;

diversity of 125, 287microgastropod grainstone 201Microphon 233Mid-Atlantic Ridge 253–54, 263millerettids 217, 220, 281millipedes 20, 130, 216, 288Mino-Tanba Belt, Japan 176Miocene 105, 151Miocidaris 189Mittiwali Member 164moas 79, 301moldavite 106; tektites 107, 108molecular biology 94, 149Molecular Clock 149

molecular phylogeny 148, 149molluscs 26, 44, 89, 105, 132, 133, 161, 166,

236, 281; bivalves 134, 191–92; Permian190; recovery of after the end-Permianextinction 193; reef 186

Mongolia 215Moores, Eldridge 254mosasaurs 89, 134Moschops 210Moschorhinus kitchingi 220mosses 216, 222, 279, 280Mount St Helens, eruption of 117mudstones 46, 61, 97, 98, 106, 162, 168, 176,

177, 179, 204, 212, 222, 229, 241, 245,247, 248, 269; black 164, 269, 270; green221, 269; red 46, 269

Mukhopadhyay, S. 260, 261Mundil, Roland 171muntjac deer 109

Murchison, Sir Roderick Impey 13, 18–19, 29,33, 35, 36–40, 37, 41, 42, 43, 44, 45, 46,48, 54, 56, 57, 60, 62–63, 66, 68–69, 70,72, 101, 124, 157, 159; and New RedSandstone 45, 52; becomes President ofthe Geological Society, London 38; namesthe Devonian System 38; names thePermian system 18–19, 35, 36, 38–39, 55,159–60, 166; names the Silurian System38; publication of note on the Permiansystem 36, 40, 51, 53; visits to Russia 18–19, 30, 36, 38, 49, 50–54, 158, 239

Muschelkalk 46mushrooms 279, 280mussels 134, 190mutation 92, 93Myosaurus 225, 291

NASA 11, 12natural selection 85, 195

Nature 115, 116nautiloids 186Nautilus 44Nemesis 139, 140neocatastrophism 13, 92, 94, 96neo-Darwinian 82, 90, 92, 93Neptune 140neutron activation 97Newell, Andy 242, 245Newell, Norman 252, 253New Red Sandstone 45–46, 49, 50, 51, 52, 53Noah’s flood see Flood, Greatnoble gases 259nonprogressionism 64, 66, 69Nopcsa von Felsö-Szilvás, Baron Franz 83–84,

83Nopcsa, Ilona 83Norell, Mark 150

Novacek, Mike 150Novikov, Igor 234nucleic acids 149; DNA 149–50; RNA 149–50nutrient cycling processes 296

Ochev, Vitaly 237, 238, 239, 242octopus 44, 132, 169, 190Oka, Russia, salt deposits 159Old Red Sandstone 45–46Olenekian: event 256–57; stage 297olivine 262Olson, Everett C. 248oolites 162Oört comet cloud 139–40Ordovician 129, 130, 131Orenburg 28, 33, 34, 51, 52, 53, 237, 239;

mines 28–29, 34, 240; Copper Sandstones33, 35, 36

orthogenesis 80, 81, 82ostracods 132, 133, 166, 173, 201Ostria acuminata 45Otocerataceae 294overkill hypothesis 79, 80Owen, Professor Richard 13, 18–19, 24, 28,

29, 30–33, 31, 34–35, 52–53, 66, 74–75,79, 207, 208, 209, 211, 214, 218; namesthe ‘Dinosauria’ 74

oxidation of coals 271oxygen 167, 281, 282, 298, 300; isotopes 172,

267, 273; levels, fall in 74, 163, 164, 169,176, 193, 250, 279; see also anoxia, lowoxygen levels and superanoxia

oyster 25, 45, 134, 190; see also Gryphaeaozone layer, destruction of 88

Pakistan 14, 163, 164, 175; evidence for end-Permian mass extinction 175, 176, 178;

Permian sections in 165palaeobiology 85, 92Palaeocene 84, 154, 273, 298Palaeontological Research Institute, Moscow

215, 227, 230–33, 234palaeosols 267–68Palaeotethys Ocean 165Palaeozoic 47–48, 48, 73, 75, 92, 133, 196,

198, 252Pander, C. H. 50Pangaea 133–34, 254, 255Panthalassa Ocean 176paper pectens 193, 201; see also Claraiaparasites 90, 289pareiasaurs 215, 218, 219, 220, 226, 230–31,

231, 233, 233, 236, 248, 291, 281Parkinsonia parkinsoni 44–45Partridge, Darren 242

Passenger pigeon 289‘patch reefs’ 295peats 268pelycosaurs 210Penny Cyclopedia 48, 49periodicity of mass extinctions 138–40Perm, Russia 13, 19, 29, 36, 52, 159; Province

28Permian: Chinese system 158–59; coals,

erosion of 270; Late 20, 23, 24, 34, 156,163, 184; naming of by Murchison 18–19,35, 36, 38–39, 55, 159–60, 166; NorthAmerican system 158–59, 160; system ofrocks in Russia 18–19, 21, 28, 33, 35, 53,157–58, 159; see also end-Permian massextinction

Permo-Triassic boundary 21, 73, 93, 156, 157,163, 165, 172, 199, 219, 227, 241–42,247, 249, 256, 258, 274; dating 164, 252;

Gartnerkofel 267; in Australia 258, 268; inAntarctica 258, 268; in Greenland 252; inHungary 259, 260; Italy 252; in Japan 259;in the Karoo Basin 206; Russia 245; SaltRange of Pakistan 257; see also end-Permian mass extinction, ‘golden spike’and Meishan section

Permo-Triassic: Britain 162; France 162;Germany 162; Greenland 223–24; rocks211; Russia 21, 229, 243; South Africa 21;the Urals 239

Petropavlova 240–41Phanerozoic eon 129, 130, 131, 132Phillips, John 46–49, 48, 55, 57, 69, 70, 154,

204, 252Philosophical Journal 39, 51Philosophical Magazine and Journal of

Science 36, 40Philosophical Transactions of the Royal

Society of London 26

phosphate 128photosynthesis 100, 101, 171, 182, 186, 280,

300phylogenetic trees 149, 150phylogenies 148, 149, 150, 151phytoplankton 101, 171PIN see Palaeontolgoical Research Institute,

Moscowpituitary gland 84, 86plagioclase feldspar 262planar deformation features 258Planet X 139, 140plankton 89, 90, 111, 129, 134, 142, 152,

181–84, 202, 280; see also foraminiferaplants: basal 150; cover, loss of 221–22, 228;

domestication of 301; extinction 101,221–24, 295; flowering 90, 130, 135;fossils 105, 162, 204, 213; groups 10;move to land 129, 130; rhyniophyte 128

plate tectonics 8, 65, 112, 221, 247, 253–54Plateosaurus 30, 292–93Pleistocene 51, 77, 78; extinctions 76, 77–78,

79–80, 134–35; mammals 79plesiosaurs 31, 74, 89Pliocene 124, 134Pluto 140poisoning 111; of the atmosphere 7, 12, 22; of

the sea 11; uranium 88polar: ice caps 65; oceans 277; regions 276;

melting 279pollen 117, 223, 224pollution 289, 302Poreda, Robert 11, 261Precambrian 128, 130, 131, 280predation 90, 138procolophonids 217, 218, 220, 225, 233, 233,

234, 236, 240, 248, 281

Procynosuchus 218productivity 171, 172, 175, 270–71, 274prolacertiform 234Promyalina 294, 294proteins 149protozoans 89, 287Protungulatum Community 110pterosaurs 74, 89, 130, 133, 134pyrite 163, 165, 168, 169, 269pyroxene 262

Qualen, F. Wangenheim von 34, 211quartz 105, 169; shocked 121, 140, 258, 262,

268, 282quartzite 241, 245, 246

racial senility 81–82, 83, 85, 87, 92radiation: cosmic 88, 92–93; UV 88, 90radioactivity, discovery of 73

radiolaria 173, 184radiolarian cherts 184radiometric dating 73, 128, 144, 170, 172, 179Rampino, Michael 185Raphanodon 233rarefaction 202rats 79, 250, 300rauisuchians 292, 293Raup, David 22, 113, 136–38, 139, 140, 152–

53, 202, 303recovery phase 10, 15–16, 153–55red beds 159, 162, 221‘red books’ of the International Union for the

Conservation of Nature 301‘reef gap’ 186, 188, 295reefs 89, 129, 132, 134, 181, 186, 281, 293;

Carboniferous period 189; diversificationof 129; end-Permian mass extinction 186;

fauna 133; formation 187, 188; frameworkof 186; poisoned 299

regression analysis 136–38, 137reptiles 13, 14, 18, 19, 21, 23, 28–29, 30, 33,

74, 81, 92, 133, 204, 206–8, 295; basal 20;diversification 256; eggs of 19; evolutionof 13, 14, 18–20, 291; fossil 13, 19, 31,92, 159, 162, 228; from CopperSandstones 28–29, 30, 33, 34, 36; killedduring KT event 152; marine reptile fossils25, 31; Mesozoic 75, 85; rate of extinctionamong 227; species loss during end-Permian 220; survival of end-permian event220

Reptilia 20Retallack, Greg 222, 258, 268, 276rhinoceros 21, 30, 76, 79, 250Rhodes, Frank 253Rhopalodon 34, 35

rhynchosaurs 291, 292rhyniophyte 128rhyolite lava 263richthofenids 191Ries Crater, Nördlingen 103–9, 104, 110, 121rift valleys 205Riley, Henry 30rivers, braided 221, 243RNA 149–50Rotliegendes 46, 49Royal College of Surgeons, London 31Royal Institution, London 37Royal Society, London 234rudists 89runaway greenhouse effect 276–77, 304Russian Academy of Sciences, Moscow 28,

230, 234Rutherford, Ernest 73

Ruzhentsev, V. E. 158Rybinsk, Russia 234Rybinskian fauna 234, 236Rybinskian Gorizont 236Rychkov, P. I. 28

sabre toothed animals 4, 21, 210–11, 218, 248,281, 291, 292

St Petersburg 29, 50, 54, 160, 215; Academyof Science 34; Mining Institute 36, 211

Sakmarian stage 158salamanders 19; modern 234, 236salt beds 52, 117, 119Salt Range mountains, Pakistan 93, 164Sambula 241, 245–46sampling 143–44sandstone 28, 33, 35, 46, 61, 97, 106, 162,

176, 216, 220, 221, 245, 247, 248; copperrich 52; green 52; marine 159; New Red

45–46, 49, 50, 51, 52, 53; Old Red 45–46;post-Carboniferous 46; red 45, 46, 52, 212,229

Saratov 237, 249; University 226Sasayama section, Japan 260, 261Sauria 18, 20saurians 24, 31, 32, 33, 53, 74Sauropoda 84Saurosternon 218Scalopognathus 236Schindewolf, Otto H. 10, 91–94, 91, 95, 96,

103, 164, 252, 257Schuchert Dal Formation 177, 178, 223Science 96, 115, 260Science Digest 112scorpions 130, 301sea level: changes in 61, 66, 71,92, 111, 113,

121, 138, 187; rise in 88, 164; fall in 88,

132, 166, 253, 254, 267, 276sea lilies see crinoidssea urchin 25, 181, 186, 189–90, 202seals 129, 184Sedgwick, Adam 38, 46, 47, 50, 56, 57, 63, 66,

68, 69, 70; names the Devonian system 38,39; names the Palaeozoic era 48

sedimentary log 243, 244sedimentation patterns 97, 98, 99, 103, 110,

161, 245: change in 221, 247, 250; rates174

seed ferns 130, 224, 280, 292; see alsoGlossopteris

Seeley, Harry Govier 208, 211selectivity 144–45, 152, 249Sennikov, Andrey 230, 234, 237Sepkoski, Jack 136–38, 139, 140, 146, 147shales 52, 177, 220, 236, 294; anoxic 294;

black 164, 168, 176, 236, 269, 294, 296

sharks 25, 129, 181, 184, 201, 280Sheehan, Peter 144Shminke, Leonid 238Shoemaker, Gene 103, 105, 106, 109, 114Shoemaker-Levy asteroid 105shrimps 174, 184, 186, 281Siberia 39, 267, 281Siberian Traps 262, 265, 271, 274–75, 275,

276, 281, 282, 304; killing model for 265–66, 282

Signor, Phillip 174Signor-Lipps effect 173–75, 185, 255silica 176, 184silicon dioxide 184Silurian 38, 41, 42, 45, 48, 49, 50, 51, 66, 69,

73, 129, 130, 131, 147; fossils 69, 262silverfish 216Simpson, George Gaylord 82

Sloan, Robert 110sloth 28, 34; giant ground 135, 196;

Megatherium 27Smith, Roger 220, 221, 222, 227, 247, 248Smith, William 43, 44, 45, 47, 56, 69snails 109, 181, 216snakes 19, 24, 109, 111, 301solar flare 8, 88Sosio Valley, Sicily 163South Africa 4, 21, 205–8, 210, 211, 212, 215,

218, 220, 225; see also Karoo Basinspecies 127, 135; conservation of 301;

diversity 14, 125, 192, 195; evolution ofnew 127; identifying new 285–87; loss140–42, 142, 180, 249–50, 300; richness126; see also biodiversity

Spencer, Patrick 242spermatogenesis 87spiders 20, 130, 132, 216, 280, 288

Spieker, Edmund 103spinescence 82Spitsbergen 267, 269sponges 186, 188, 295squid 44, 132, 169, 190squirrels 250Staithes, Yorkshire 25Stanley, Steven 256, 266starfish 186, 189Stegosaurus 82Stephanoceras humphriesianum 44Stevns Klint section, Denmark 99, 100Stormberg Group 213Storrs, Glenn 230, 234, 237, 242stratigraphic consistency index (SCI) 151stratigraphy 38, 42, 43, 44, 45, 46, 47, 52, 69,

70, 73, 91, 92, 111, 148, 149, 150Streptognathodus isolatus 162

stromatoporoids 133strontium ratios 276Stutchbury, Samuel 30suevite 104, 105, 107, 109, 119, 120sulphate gases 265, 266; cooling effect of

275–76sulphur 163; dioxide 121, 265, 266, 275, 277sulphuric acid 276, 279sun, blocked out by dust 100, 101, 108, 111sunspots 71, 88superanoxia 269, 270, 271superfamilies, extinction of 191supernova 92, 94, 114, 121Surkov, Mikhail 226synapsids 20–21, 22, 23, 34, 35, 207, 209–10,

211, 212, 218, 231, 233, 238, 256, 291synonymy 285–86Syodon 34

tadpoles 19taphonomy 144–45, 213Tapinocephalus Zone 256Tatarian stage 157, 162Tatarinov, Leonid P. 227taxonomic hierarchy 126, 141Tchijevski, A. L. 194, 198Tchudinov, Peter 238, 239tectonic activity 221; see also plate tectonicsTeichert, Carl 156, 163–64, 165; ‘mudball

theory’ 177–78tektites 97temnospondyl 233, 234–36, 239temperature change 113; fall in 9, 12, 22, 77,

85, 87, 88, 90, 101, 111, 138, 153; rise in12, 57, 77, 87, 88, 90, 172, 175, 222, 250,266, 267, 277, 280, 281; see also climatechange

Tertiary 89, 119–20, 129, 131, 134Tesero Oolite Horizon 162, 163Tethys Ocean 175, 181, 256, 293tetrapods 20, 23, 146, 153–54Thecodontosaurus 18, 30therapsids 210, 218therocephalian 218, 220, 225, 231, 233, 234,

248, 281Thoosuchus 236Thrinaxodon 210tidal wave see tsunamiTikhvinsk 233, 234, 236, 240, 241Tikhvinskia 234, 236tillites 267, 268Toishi Shale 176tortoises 109, 111Triassic 18, 23, 30, 35, 46, 49, 51, 52, 53, 80,

92, 93, 129, 131, 176, 193, 199, 296; base

of 162–63; Early 23, 24, 164, 176, 181,236, 256, 266, 267, 268, 269, 270, 297,304; fish groups 200; fossils of 76, 169,175; Jurassic boundary 133; Late 24, 136;Mid 264; mudstones 106; of Germany 46,49 74; recovery of life during 256; reefs188; rocks 18, 19, 35, 204; sandstones106; sediments 108

Triceratops 102trilobites 54, 66, 75, 92, 92, 128, 132, 133,

150, 198–99; fossils 173tsunamis 111, 118, 119, 121tuff 169, 265, 266Tunguska, Siberia 103turtles 21, 31, 32, 95, 133, 150, 209, 217Tverdokhlebov, Valentin 229, 237, 238, 239,

240, 241–42, 243, 249Twelvetrees, William Harper 211Twitchett, Richard 178, 223, 224, 269

typostrophism 91–93Tyrannosaurus rex 102, 293

Ufimian stage 157ultraviolet radiation 88UNESCO 160–61uniformitarianism 58–62, 63, 64, 65, 80uniformitarians 9, 57, 68‘uniformity’ 64–66Unionites 294, 294Ural Mountains, Russia 12–13, 19, 28, 29, 33,

35, 51, 52, 54, 159, 162, 213, 229, 231,237, 240, 241, 245, 246, 247; uplift of 254

uranium 239; poisoning 88uranium-lead dating method 170

Valentine, Jim 252, 254Van Valen, Leigh 110, 112Vendian period 128, 129, 130, 131

Verneuil, Edouard de 50, 51–52, 54Vinci, Leonardo da 25viruses 287vis plastica 26Vokhmian Gorizont 236volcanic: ash 99, 104; dust 9, 87, 99–100;

eruptions 7, 14, 22, 59, 61, 64, 66, 85, 101,117, 121, 123, 135, 262, 277–79, 282;fragments 169; lava 63, 104, 117, 119,121, 206, 213, 262, 263, 277, 281, 303;vent 265; see also basalt, lava, volcanoesand vulcanism

volcanoes 11, 12, 27, 63, 99, 103, 104, 119,123, 266, 277; conical 263

Voltzia 268Voltziopsis 268vulcanism 7, 65, 87, 88, 92; see also volcanic

eruptions, volcanoesVyatskian assemblage 233, 233, 236

Waagen, W. 164Waldheim, J. G. F. Fischer von 34, 35, 36, 39,

50, 53, 211Wallace, Alfred Russel 79Ward, Peter 147, 220, 221, 222, 227, 247, 248Watson, D. M. S. 93Wegener, Alfred 253, 254, 255Wetlugasaurus 234whales 129, 184Wheeler, Raymond P. 194Whitby, Yorkshire 25, 26Wignall, Paul 164, 166, 169, 175, 178, 201,

266, 267, 269, 271, 274Wills, Matthew 150Woodward, Arthur Smith 81, 82Wooler, Mr 26woolly rhinos 76, 135, 196, 219Wordie Creek Formation 177, 178, 179, 223

Xenodiscidae 294

Yang, X. 256Yarenskian Gorizont 241Youngina 218

Zechstein 34, 43, 46, 49, 52, 53, 162Zumaya, Spain, KT boundary 147Zygosaurus 34

Originally published in the United Kingdom in2003 as

When Life Nearly Died: The Greatest MassExtinction Of All Time

ISBN 978-0-500-05116-0by Thames & Hudson Ltd, 181a High Holborn,

London WC1V 7QX

Copyright © 2003 and 2008 Thames & HudsonLtd, London

This electronic version first published in 2013by

Thames & Hudson Ltd, 181a High Holborn,London WC1V 7QX

First published in 2013 in the United States ofAmerica by

Thames & Hudson Inc., 500 Fifth Avenue, NewYork, New York 10110

To find out about all our publications, pleasevisit

www.thamesandhudson.comwww.thamesandhudsonusa.com

All Rights Reserved. No part of this publicationmay be reproduced or transmitted in any form

or by any means, electronic or mechanical,including photocopy, recording or any other

information storage and retrieval system,without prior permission in writing from the

publisher.

ISBN 978-0-500-77100-6ISBN for USA only 978-0-500-77101-3

On the cover: Photo Mark Westerby ©Untitled Photo Library