the mother of mass extinctions

72 Scientific American July 1996

The Mother of Mass ExtinctionsDisaster struck 250 million years ago, when the worst decimation in the earth’s history occurred. Called the end-Permian mass extinction, it marks a fundamental change in the development of life

by Douglas H. Erwin

PATR

ICIA

J. W

YN

NE

1 DROP IN SEA LEVEL, which had begungradually about 260 million years ago, be-came quite sudden at the very end of thePermian, destroying near-shore habitatsand destabilizing the climate.

2 INCREASED OXIDATION oforganic matter raised levels ofcarbon dioxide in the atmo-sphere, contributing to globalwarming and depleting theamount of oxygen that coulddissolve in water.

FORAMINIFERA

RUGOSE CORALS

Copyright 1996 Scientific American, Inc.

The history of life on the earth is replete with catastrophes of varying magnitudes.The one that has captured the most attention is the extinction of the dinosaursand other organisms 65 million years ago—between the Cretaceous and Tertiary

periods—which claimed up to half of all species. As severe as that devastation was, it palesin comparison to the greatest disaster of them all: the mass extinction, some 250 millionyears ago, at the end of the Permian period. Affectionately called “the mother of mass ex-tinctions” among paleontologists (with apologies to Saddam Hussein), it yielded a deathtoll that is truly staggering. About 90 percent of all species in the oceans disappeared duringthe last several million years of the Permian. On land, more than two thirds of reptile andamphibian families vanished. Insects, too, did not escape the carnage: 30 percent of insectorders ceased to exist, marking the only mass extinction insects have ever undergone.

But from catastrophes, opportunities arise. For several hundred million years before theend-Permian event, the shallow seas had been dominated by life-forms that were primar-ily immobile. Most marine animals lay on the seafloor or were attached to it by stalks,filtering the water for food or waiting for prey. In the aftermath of the extinction, manyonce minor groups—active, predatory relatives of modern-day fish, squids, snails andcrabs—were able to expand. Some completely new lineages appeared. This ecological re-organization was so dramatic that it forms a fundamental boundary in the history of life.Not only does it demarcate the Permian and Triassic periods, it also establishes the closeof the Paleozoic era and the start of the Mesozoic era. The modern tidal pool reflects

Scientific American July 1996 73

DEADLY CATASTROPHES combined to wipe out most of life on the earth at the end of thePermian period, about 250 million years ago. Reef and shallow-water communities, consist-ing of corals, sea lilies, foraminifera and other organisms, were especially hard hit. On land,more than two thirds of reptiles and amphibians and nearly one third of insects disappeared.An increase in fungal spores suggests that plants, too, suffered.

3 MASSIVE VOLCANIC ERUPTIONSthat started about 255 million yearsago and persisted for a few millionyears cooled the earth in the shortterm and led to long-term globalwarming and ozone depletion.

4 RETURN OF THE SEAshortly after the lowestlevels had been reacheddisrupted coastal com-munities and flooded in-land areas with watersthat may have beenstagnant.

FUNGALSPORES

SHORT-TERMCOOLING

LONG-TERMWARMING


what lived and what died 250 millionyears ago.

Over the past few years, exciting newinsights into the causes and conse-quences of the end-Permian mass ex-tinction have poured in from virtuallyevery branch of the earth sciences.Some of these findings include detailedstudies of rapid changes in ocean chem-istry, more thorough documentation ofextinction patterns and new analysesshowing that large volcanic eruptionsoccurred at the Permo-Triassic bound-ary [see “Large Igneous Provinces,” byMillard F. Coffin and Olav Eldholm;Scientific American, October 1993].

My own research during the past de-cade has been driven by curiosity aboutthe events that structure life’s evolution-ary patterns. How much do mass ex-tinctions contribute to the evolution of

a group, as compared with long-termadaptive trends? For example, sea ur-chins are ubiquitous in modern oceansbut were relatively uncommon duringthe Permian. Only a single genus, Mio-cidaris, is known for certain to havesurvived the extinction. Did Miocidarissurvive by pure chance, or was it betteradapted? Would sea urchins today lookany different had it not been for theend-Permian extinction?

A Few Good Rocks

To resolve such questions, we needto learn more about the causes of

the catastrophe and how those speciesthat survived differed from those thatdisappeared. The key sources for thisinformation are rock layers and fossils.Unfortunately, samples from the late

Permian and early Triassic are notori-ously difficult to come by. The fossilrecord across the boundary is plaguedby poor preservation, a lack of rock tosample and other problems, includingaccess. An extensive drop in sea levelduring the late Permian limited thenumber of marine rocks deposited onland, and many areas where the bestrocks were preserved (most notably, insouthern China) have been relativelyhard for some geologists to reach.

As such, it has proved difficult to as-certain just how quickly life was snuffedout or if the deaths were subject to anyregional variations. Some creatures, es-pecially those sensitive to changes in theenvironment, died off rapidly, as shownby Erik Flügel and his colleagues at theUniversity of Erlangen, who arrived atthis conclusion after examining reefs insouthern China and Greece. Other evi-dence indicates more gradual loss oflife. For example, in studying the incred-ibly diverse and beautifully preservedfauna in the limestone outcroppings ofwestern Texas and adjacent New Mexi-co and Arizona, I have found that manysnails began vanishing late in the mid-dle of the Permian, well before the mainpulse of extinction.

Intensive studies of newly found andcritical boundary layers in Italy, Austriaand southern China have helped our un-derstanding. They indicate that the du-ration of the extinction is shorter thanpreviously thought, implying that ab-ruptly calamitous environmental condi-tions must have set in. Only a few yearsago, I believed the extinction period mayhave persisted five to 10 million years.It now appears that the final pulse mayhave lasted less than one million years.

Steven M. Stanley of Johns HopkinsUniversity theorizes that the extinctionmay have consisted of two brief episodes,one occurring at the end of the middlePermian and the second at the end ofthe late Permian. Jin Yugan of the Nan-jing Institute of Geology and Palaeon-tology, Samuel A. Bowring of the Mas-sachusetts Institute of Technology and Iare collaborating on a project to datevolcanic ash beds in southern China andsoon should have a better sense of thelength of the extinction period. In anycase, the rate appears to be about as rap-id as many other mass extinctions.

In this geologically brief interval, thePermian oceans experienced a complexpattern of life and death. Quantifyingthe taxonomic extent of disappearance—

from order to family to genus to spe-

The Mother of Mass Extinctions74 Scientific American July 1996

DEGRADED ENVIRONMENT during the late Permian is revealed by geochemicaland fossil evidence. At that time, more carbon was oxidized, the sea began dropping,and volcanism in what is now Siberia and China took place. Some layers in the oceansmay also have become anoxic. Reefs did not recover fully until the middle Triassic.

LATE PERMIAN

GUADALUPIANDJULFIAN

TATARIAN

CHANGXINGIAN

GRIES-BACHIAN

DIE-NERIAN

SMITHIAN/SPATHIAN

EARLY TRIASSIC

500

257 + 3– –251 243 + 3

400 300AGE (MILLIONS OF YEARS AGO)

PALEOZOICC

AM

BR

IAN

OR

DO

VIC

IAN

SIL

UR

IAN

DEV

ON

IAN

CA

RB

ON

IFER

OU

S

PER

MIA

N

TRIA

SSIC

JUR

ASSIC

CR

ETA

CEO

US

TER

TIA

RY

MESOZOIC CENOZOIC

200 100 65251540

DURATION OF ANOXIA

DURATION OF REEF RECOVERY

FORMATION OF SIBERIAN TRAPS

SOUTHERN CHINA VOLCANISM

SEA-LEVEL CHANGES

RA

TIO

OF

CA

RB

ON

12

TO C

AR

BO

N 1

3LA

ND

SU

BM

ERG

ED(P

ERC

ENT)

30

20

10

0

OXIDATION OF CARBON

JOH

NN

Y J

OH

NSO

N


cies—can be difficult. It is believed thaton a global scale, 49 percent of familiesand 72 percent of genera were wipedout. Because species are harder to iden-tify, that loss is far more difficult to pindown, and researchers have come upwith varying estimates. Yin Hongfu andhis colleagues at the China University ofGeosciences examined the rock layer de-marcating the Changxing stage in south-ern China. (This stage, along with theDjulfian stage, marks the later of the twosubdivisions of the late Permian; eachstage is named for the part of the worldwhere the fossil record is clearest.)

Yin and his colleagues reported thatout of 476 late Permian invertebrate spe-cies, 435 (or 91 percent) vanished. (Oth-er estimates of global species extinctionrange from 80 to 95 percent, but thelower end of this range is probably mostrealistic.) By way of comparison, theevent that occurred at the end of the Or-dovician period, 439 million years ago,eliminated 57 percent of marine genera.The Cretaceous-Tertiary extinction,which killed off the dinosaurs, claimedup to 47 percent of existing genera.

The end-Permian devastation hit someanimals harder than others. Groups thatlived attached to the seafloor and filteredorganic material from the water fornourishment suffered the greatest ex-tinction. They included corals, articu-late brachiopods (a kind of shelled in-vertebrate), some bryozoans (filter feed-ers that lived in colonies) and a varietyof echinoderms (sea lilies). Other deci-mated marine groups included the last

few trilobites, shallow-water foraminif-era (a type of zooplankton) and ammo-noids (distant relatives of the nautilus).Snails, bivalves and nautiloids camethrough the period fairly well, sufferingjust a few group losses. The only ma-rine group that was truly indifferent tothe mounting chaos was the conodonts,primitive chordates whose easily pre-served mouthparts serve as importantmarkers of time.

Things were not much better on land.Terrestrial vertebrates and insects bothexperienced substantial losses. Amongvertebrates 78 percent of reptile and 67percent of amphibian families disap-peared during the late Permian, al-though how rapidly this occurred re-mains a subject of debate. Earlier stud-ies from the magnificent fossils found inthe Karroo region of South Africa sug-gested the decline took place over sever-al million years, perhaps with two peaksin the extinction rate. Some recent work,though, suggests a more rapid decline,similar to the pace of marine extinction.

The extinction of insect species marksa major transformation of fauna. Ofthe 27 orders of insects known from thePermian, eight became extinct near thePermo-Triassic boundary, four sufferedsevere decimation but recovered, andthree more barely survived into the Tri-assic before becoming extinct. This isthe only significant insect extinctionevent yet identified, and it serves as atestament to the severity of the environ-ment at the time.

Terrestrial flora suffered as well. To

what extent, however, is impossible tosay, for the evidence on the magnitudeis, at the moment, not especially solid. Inexamining Australian leaf fossils, GregJ. Retallack of the University of Oregonshowed this past year that plant extinc-tions were far more dramatic than hadbeen thought and led to a rapid shift inthe dominant floral types in Australia.(The loss of plant life may have alsocontributed to the disappearance of in-sects that fed on the flora.)

The record of pollen and spores moreaccurately reflects the effects on plants.In late Permian strata, pollen from gym-nosperms (woody plants such as coni-fers) is almost absent, and succeedinglayers harbor only fungal cells and someadditional organic detritus. Last yearHenk Visscher and his colleagues at theUniversity of Utrecht in the Netherlandsfound that this so-called fungal spikeseems to have begun in a latter part ofthe Permian—specifically, in the lateChangxingian stage—before reachingits climax at the Permo-Triassic border.

Extinction Traps

Given the evidence of marine andterrestrial fossils, clearly the late

Permian was a period when almost ev-erything went wrong—at least if a spe-cies wanted to stay alive. What couldhave caused the wholesale loss of life?About the only thing that did not hap-pen, or at least for which we have noevidence, was an extraterrestrial impact,an event that most likely killed off the

The Mother of Mass Extinctions Scientific American July 1996 75

PERMIAN LIFE recorded in fossilized remains includes the bry-ozoans (left) and the brachiopods (right), two closely related

phyla that were major components of marine life before the ex-tinction. Each fossil is about 40 millimeters long.

KATH

ERIN

E LA

MB

ERT


dinosaurs. In the mid-1980s a group ofgeologists claimed to have found vanish-ingly small traces of iridium, a criticalindicator of a collision, at the Permo-Triassic boundary layer in southernChina. But despite many attempts, noone has substantiated these assertions.

There is no shortage of murder sus-pects, however. One possibility is volcan-ism. A key piece of evidence is the Siber-ian traps, solidified layers of ancient lava.The traps (after the Swedish word for“stairs,” which describes the steplikeedges of the deposits) include at least 45separate flows and range from 400 to3,700 meters in thickness. They cover atleast 1.5 million cubic kilometers, andperhaps more, for they may extend west-ward under younger rocks to the UralMountains. (In comparison, the eruptionof Mount Pinatubo in 1991 was a merepuff, one that spewed ash but no mag-ma. Perhaps a better comparison is theLaki eruption in Iceland, which in 1783produced 15 cubic kilometers of lava.)

Recent radiometric dating suggeststhat all the lava of the Siberian traps mayhave erupted within a period of less thanone million years, and perhaps in only600,000 years, beginning near the Per-mo-Triassic boundary and extendinginto the earliest Triassic. Paul R. Renneof the Berkeley Geochronology Centerhas found that this estimate matcheswell with other large eruptions that de-posited volcanic ash in southern China.

Could extensive volcanism havepurged the earth during the late Permi-an? Eruptions have a variety of short-term effects, including cooling from

both dust and sulfates ejected into thestratosphere (remember Pinatubo), acidrain, wildfires, release of poisonous traceelements and an increase in ultravioletradiation from ozone-layer depletion.And over a longer timescale, the carbondioxide emitted may lead to warming.

As appealing as this hypothesis is,killing some 90 percent of the species inthe oceans is remarkably difficult. By it-self, volcanism, even as rapid and ex-tensive as that which produced theSiberian traps, is not up to the task. Mycolleague Thomas A. Vogel of MichiganState University and I examined volcanicash sheets deposited by eruptions duringthe past 100 million years. These erup-tions were similar in magnitude to thosethat produced the ash in southern Chi-na at the end of the Permian. We foundthat none of these events greatly affect-ed the diversity of regional and globallife on land or in the oceans.

Moreover, the environmental damageproduced by any eruption depends onseveral factors. Many volcanic effects,such as the amount of sulfate ejected intothe stratosphere, are difficult to inferfrom eruptions that occurred 250 mil-lion years ago. So eruptions may havebeen involved in the extinctions, but onlyas part of a more complex process.

Geochemistry and the Stagnant Sea

The most intriguing new evidenceabout the end-Permian mass extinc-

tion comes from the field of geochemis-try. Perhaps the most relevant geochem-ical changes are the shifts in carbon iso-

topes found in rocks (specifically, theratio of carbon 12 to carbon 13). Thisfact indicates that, apparently, more or-ganic matter was being buried duringthe late Permian than in previous times.

Although this burial of carbon is tell-ing us something about geochemicalchanges during the end-Permian extinc-tion, it is not entirely clear what. It mayhave to do with the sudden, deadly dropin sea level. During the early Permian,the continents merged to form the sin-gle supercontinent Pangaea. Around thecontinental shelves, reefs and other shal-low-water communities thrived. Then,near the end of the Permian, the sea lev-el fell. (No one knows exactly why, butit may have been caused by changes inthe earth’s mantle that enlarged theocean basins.) The drop disrupted thehabitats along the shore. With more ofPangaea’s continental shelf exposed,greater erosion and oxidation of organicmatter probably occurred. This oxida-tion reduced the oxygen and increasedthe carbon dioxide in the atmosphere,which may have humidified the planetand warmed it by as much as two de-grees Celsius.

Further disruption occurred when sealevels rose again, perhaps several hun-dred thousand years later. The risingocean waters engulfed near-shore habi-tats and swept inland. Such intrusionsundoubtedly killed off many coastalcommunities.

Decreased amounts of atmosphericoxygen might also have exacerbated thehostile conditions already developing.Less oxygen would have dissolved into


EXTINCTION VICTIMS include fauna from the Karroo re-gion of South Africa, where this 36-centimeter-long Oudendon

(left), a therapsid (mammallike reptile), was recovered. Thenine-centimeter-wide ammonoid (right) was found in Texas.

KATH

ERIN

E LA

MB

ERT


the oceans, causing anoxia, which couldhave suffocated some marine life. Theevidence for anoxic waters is reflected bygeochemical anomalies. Several research-ers have recently made the intriguing,though not completely compelling, argu-ment that patterns of extinction amongvarious species reflect the abilities of theorganisms to withstand anoxia.

What actually caused the extinctions?There may not be a single smoking gun,but rather all the possibilities mentionedmay have contributed. None of themalone would have caused an extinctionof this size, but it was the bad luck ofthe exquisite Permian faunas to have allof them interact at about the same time.

I believe the extinction consisted ofthree phases. The first began with thedrop in sea level around much of Pan-gaea, which led to a loss of habitat, cli-matic instability and the elimination ofmany narrowly distributed species. As

the oceanic regression continued, phasetwo began, with volcanic eruptions andthe release into the atmosphere of largevolumes of carbon dioxide, which in-creased climatic instability and facilitat-ed ecological collapse. The rise in sealevel and subsequent floods of possiblyanoxic waters at the very end of the Per-mian and into the early Triassic initiatedthe third phase. It destroyed near-shoreterrestrial habitats and contributed tothe extinction of many surviving taxa.

Life after Death

The aftermath of the end-Permianextinction is at least as interesting as

the event itself. After other mass extinc-tions, life began to recover within abouta million years. In this case, though, ittook perhaps five million years. The lagstems from the fact that biological com-munities had been so severely disrupted

that millions of years were required forthem to re-form and flourish. (There isalso the possibility that the recovery ap-pears longer than it actually was becauseof the poor preservation of fossils.)

Regardless of how long it took to be-come reestablished, life on the earth hadchanged dramatically. As I noted earlier,largely immobile animals had dominat-ed the Permian seas: brachiopods, bryo-zoans and echinoderms. They sat on thebottom, filtering the water for food orwaiting for prey to swim by. The mobileanimals—fish, bivalves, cephalopods(squids and their relatives) and gastro-pods (snails)—were around but formedonly a small part of the community. Afew trilobites remained.

Shortly after the extinction, duringwhat is called the survival stage of theearly Triassic, the few remaining speciestended to be abundant and widespread.Earliest Triassic faunas consist of some

The Mother of Mass Extinctions Scientific American July 1996 77

PERMIAN PERIOD1 SPONGE

2 CRINOID

3 BRACHIOPOD

4 NAUTILOID

5 BEADED SPONGE

6 BRYOZOAN

7 CORAL

8 TRILOBITE

9 ALGAE

10 SNAIL

11 FISH (Janessa)

12 FISH (Dorypterus)

CRETACEOUS PERIOD1 COELACANTH

2 AMMONOID

3 BELEMNOID

4 SNAIL

5 BIVALVES (RUDISTS)

6 SEA URCHIN

7 CRAB

8 ALGAE

9 SCALLOP

10 FISH (Thrissops)

11 FISH (Davichthys)

12 STARFISH

STRUCTURE OF MARINE LIFE was dramatically changedafter the mass extinction. In the middle Permian (left), the seascontained mostly immobile animals, with some fish and a few

trilobites. But by the Cretaceous period (right), the ocean re-sembled modern-day seas, with mobile bivalves, gastropods,swimming fish and cephalopods.

PATR

ICIA

J. W

YN

NE

PATR

ICIA

J. W

YN

NE 4

2

3

1

6

7

8

9

10

11

121

2

3

4

2

5

6

7

12

9

8

10

3

2

4

411

2

2

5


clams, ammonoids and a few gastro-pods. The reptilelike Lystrosaurus, anancestor to mammals, was the mostcommon vertebrate on land and wasfound throughout Pangaea. The clamClaraia thrived in the oceans. By themiddle of the Triassic, though, some 25million years later, sea ur-chins and other groups thatwere relatively sensitive totheir environments began toreemerge, marking the startof a return to more normalmarine conditions. These“Lazarus taxa” (as David Ja-blonski of the University ofChicago refers to them, fortheir return from the dead)began displacing the surviv-ing fauna.

Indeed, by this time theseas were bearing some re-semblance to the modernoceans. The more mobilecreatures, such as bivalves,gastropods and crabs, dominated. Agreater diversity of cephalopods andother predatory swimmers also emerged.Burrowing became fashionable, perhapsreflecting a need for more protectionfrom predation. An evolutionary armsrace between predator and prey ensued,driving changes in skeletal architectureon through the Mesozoic era that yield-ed fauna that had more flesh than Pale-ozoic creatures had. Such changes pro-duced more complex and sophisticatedecosystems—there was more to eat andmore of a menu selection, too.

Details of evolutionary changes onland for the same period are still some-what sketchy, because the fossils haveyet to receive a detailed, bed-by-bed sam-pling. Those proposed studies promiseto boost our understanding of the terres-trial extinctions. We do know that sev-eral reptile and amphibian groups came

to an end. Also, insects shifted from avariety of dragonflylike groups, whichhave wings that were fixed in the flightposition and that could not be foldedover the body, to forms that could foldtheir wings. These newer forms, whichmake up 98 percent of the insects of to-

day, also had separate larval and adultstages. The adaptations may reflect anability to exploit new habitats and towithstand severe seasonal swings andother climatic instabilities.

Survival of the Fittest?

The changes that took place amonginsects raise the general question of

whether species that made it successful-ly out of the Permian had specific adap-tations that enabled them to survive orwhether their survival was more random.Fossils of Claraia are found in rocks thatharbor evidence of anoxic conditions.The large numbers and widespread dis-tribution of Claraia may indicate thatthis species could survive on little oxy-gen. Another example is Miocidaris, thesole echinoid (sea urchin) survivor (al-though a close relative very likely sur-

vived as well). Miocidaris has only twocolumns of interambulacral plates(roughly speaking, the areas between the“petals” on the shell of a sea urchin);other Permian echinoids had anywherefrom one to eight columns of plates. Be-cause Miocidaris was the only genus to

survive, the predominantform of echinoids shiftedfrom those that had highlyvariable numbers of columnplates to those that have onlytwo. Some paleontologistshave argued that an echinoidskeleton is stronger if it iscomposed of only two col-umns of plates and, thus, per-haps better adapted to sur-vive predation in the post-Permian world.

Unfortunately, it is nearlyimpossible to say whether thePermo-Triassic extinction se-lected for certain features.All modern echinoids might

have developed two-column plates evenif the end-Permian extinction had neveroccurred. The surviving fauna may sim-ply consist of groups that were the mostabundant and widely distributed beforethe extinction and thus had the bestchance to survive. Distinguishing be-tween these two possibilities has turnedout to be quite tough.

The only thing we can say for certainis that the end-Permian mass extinctionhad the greatest effect on the history oflife of any event since the appearance ofcomplex animals. Without this episode,there is little doubt that the composi-tion of a modern tidal pool would lookvastly different. Children would havegrown up learning about crinoids andbrachiopods instead of starfish and seaurchins, perhaps even looking in poolsto catch a fleeting glimpse of a passingtrilobite.


The Author

DOUGLAS H. ERWIN is a research paleobiologist and curator of Paleo-zoic gastropods and of the Burgess Shale in the department of paleobiologyat the National Museum of Natural History of the Smithsonian Institution.His research focuses on the evolutionary history of Paleozoic gastropodsand on large-scale evolutionary patterns, particularly the explosive spread ofanimals during the Cambrian and after the end-Permian mass extinction. Hehas studied Cambrian and Permian rocks in Siberia, China, Newfoundlandand throughout the western U.S. In addition to one solo effort, he wrotewith Derek Briggs and Fred Collier Fossils of the Burgess Shale (Smithsoni-an Press, 1994) and edited with Robert Anstey New Approaches to Specia-tion in the Fossil Record (Columbia University Press, 1995).

Further Reading

Extinction: Bad Genes or Bad Luck? David M. Raup. W.W. Norton, 1991.

Permo-Triassic Events in the Eastern Tethys: Stratig-raphy, Classification, and Relations with the West-ern Tethys. Edited by Walter C. Sweet, Yang Zunyi, J. M.Dickins and Yin Hongfu. Cambridge University Press, 1992.

The Great Paleozoic Crisis: Life and Death in thePermian. Douglas H. Erwin. Columbia University Press,1993.

The Permo-Triassic Extinction. Douglas H. Erwin inNature, Vol. 367, pages 231–236; January 20, 1994.

PERMIAN SEA URCHIN Miocidaris, four centimeters long,was the only genus of echinoid to have survived the extinction.

DO

UG

LAS H

. ER

WIN

SA


the mother of mass extinctions

Documents