Has the Earth’s Sixth Mass Extinction Already Arrived? Nature

Abstract

Palaeontologists characterize mass extinctions as times when the Earth loses more than three-quarters of its species in a geologically short interval, as has happened only five times in the past 540 million years or so. Biologists now suggest that a sixth mass extinction may be under way, given the known species losses over the past few centuries and millennia. Here we review how differences between fossil and modern data and the addition of recently available palaeontological information influence our understanding of the current extinction crisis. Our results confirm that current extinction rates are higher than would be expected from the fossil record, highlighting the need for effective conservation measures.

Full text

REVIEW doi:10.1038/nature09678
Has the Earth’s sixth mass extinction
already arrived?
Anthony D. Barnosky
1,2,3
, Nicholas Matzke
1
, Susumu Tomiya
1,2,3
, Guinevere O. U. Wogan
1,3
, Brian Swartz
1,2
, Tiago B. Quental
1,2
{,
Charles Marshall
1,2
, Jenny L. McGuire
1,2,3
{, Emily L. Lindsey
1,2
, Kaitlin C. Maguire
1,2
, Ben Mersey
1,4
& Elizabeth A. Ferrer
1,2
Palaeontologists characterize mass extinctions as times when the Earth loses more than three-quarters of its species in a
geologically short interval, as has happened only five times in the past 540 million years or so. Biologists now suggest that a
sixth mass extinction may be under way, given the known species losses over the past few centuries and millennia. Here
we review how differences between fossil and modern data and the addition of recently available palaeontological
information influence our understanding of the current extinction crisis. Our results confirm that current extinction
rates are higher than would be expected from the fossil record, highlighting the need for effective conservation measures.
Of the four billion species estimated to have evolved on the Earth
over the last 3.5 billion years, some 99% are gone
1
. That shows
how very common extinction is, but normally it is balanced by
speciation. The balance wavers such that at several times in life’s history
extinction rates appear somewhat elevated, but only five times qualify
for ‘mass extinction’ status: near the end of the Ordovician, Devonian,
Permian, Triassic and Cretaceous Periods
2,3
. These are the ‘Big Five’
mass extinctions (two are technically ‘mass depletions’)
4
. Different
causes are thought to have precipitated the extinctions (Table 1), and
the extent of each extinction above the background level varies depend-
ing on analytical technique
4,5
, but they all stand out in having extinction
rates spiking higher than in any other geological interval of the last ,540
million years
3
and exhibiting a loss of over 75% of estimated species
2
.
Increasingly, scientists are recognizing modern extinctionsof species
6,7
and populations
8,9
. Documented numbers are likely to be serious under-
estimates, because most species have not yet been formally described
10,11
.
Such observations suggest that humans are now causing the sixth mass
extinction
10,12–17
, through co-opting resources, fragmenting habitats,
introducing non-native species, spreading pathogens, killing species
directly, and changing global climate
10,12–20
. If so, recovery of biodiversity
will not occur on any timeframe meaningful to people: evolution of new
species typically takes at least hundreds of thousands of years
21,22
, and
recovery from mass extinction episodes probably occurs on timescales
encompassing millions of years
5,23
.
Although there are many definitions of mass extinction and grada-
tions of extinction intensity
4,5
, here we take a conservative approach to
assessing the seriousness of the ongoing extinction crisis, by setting a
high bar for recognizing mass extinction, that is, the extreme diversity
loss that characterized the very unusual Big Five (Table 1). We find that
the Earth could reach that extreme within just a few centuries if current
threats to many species are not alleviated.
Data disparities
Only certain kinds of taxa (primarily those with fossilizable hard parts)
and a restricted subset of the Earth’s biomes (generally in temperate
latitudes) have data sufficient for direct fossil-to-modern comparisons
1
Department of Integrative Biology, University of California, Berkeley, California 94720, USA.
2
University of California Museum of Paleontology, California, USA.
3
University of California Museum of
Vertebrate Zoology, California, USA.
4
Human Evolution Research Center, California, USA. {Present addresses: Departamento de Ecologia, Universidade de Sa
˜o Paulo (USP), Sa
˜o Paulo, Brazil (T.B.Q.);
National Evolutionary Synthesis Center, 2024 W. Main Street, Suite A200, Durham, North Carolina 27705, USA (J.L.M.).
Table 1
|
The ‘Big Five’ mass extinction events
Event Proposed causes
The Ordovician event
64–66
ended ,443 Myr ago; within 3.3 to
1.9 Myr 57% of genera were lost, an estimated 86% of species.
Onset of alternating glacial and interglacial episodes; repeated marine transgressions and
regressions. Uplift and weathering of the Appalachians affecting atmospheric and ocean chemistry.
Sequestration of CO
2
.
The Devonian event
4,64,67–70
ended ,359 Myr ago; within 29 to
2 Myr 35% of genera were lost, an estimated 75% of species.
Global cooling (followed by global warming), possibly tied to the diversification of land plants, with
associated weathering, paedogenesis, and the drawdown of global CO
2
. Evidence for widespread
deep-water anoxia and the spread of anoxic waters by transgressions. Timing and importance of
bolide impacts still debated.
The Permian event
54,71–73
ended ,251 Myr ago; within
2.8 Myr to 160Kyr 56% of genera were lost, an estimated
96% of species.
Siberian volcanism. Global warming. Spread of deep marine anoxic waters. Elevated H
2
SandCO
2
concentrations in both marine and terrestrial realms. Ocean acidification. Evidence for a bolide
impact still debated.
The Triassic event
74,75
ended ,200 Myr ago; within 8.3Myr
to 600 Kyr 47% of genera were lost, an estimated 80% of
species.
Activity in the Central Atlantic Magmatic Province (CAMP) thought to have elevated atmospheric
CO
2
levels, which increased global temperatures and led to a calcification crisis in the world oceans.
The Cretaceous event
58–60,76–79
ended ,65 Myr ago; within
2.5 Myr to less than a year 40% of genera were lost, an
estimated 76% of species.
A bolide impact in the Yucata
´n is thought to have led to a global cataclysm and caused rapid cooling.
Preceding the impact, biota may have been declining owing to a variety of causes: Deccan
volcanism contemporaneous with global warming; tectonic uplift altering biogeography and
accelerating erosion, potentially contributing to ocean eutrophication and anoxic episodes. CO
2
spike just before extinction, drop during extinction.
Myr, million years. Kyr, thousand years.
3MARCH2011|VOL471|NATURE| 51
Macmillan Publishers Limited. All rights reserved
©2011

(Box 1). Fossils are widely acknowledged to be a biased and incomplete
sample of past species, but modern data also have important biases that,
if not accounted for, can influence global extinction estimates. Only a
tiny fraction (,2.7%) of the approximately 1.9 million named, extant
species have been formally evaluated for extinction status by the
International Union for Conservation of Nature (IUCN). These IUCN
compilations are the best available, but evaluated species represent just a
few twigs plucked from the enormous number of branches that compose
the tree of life. Even for clades recorded as 100% evaluated, many species
still fall into the Data Deficient (DD) category
24
. Also relevant is that not
all of the partially evaluated clades have had their species sampled in the
same way: some are randomly subsampled
25
, and others are evaluated as
opportunity arises or because threats seem apparent. Despite the limita-
tions of both the fossil and modern records, by working around the
diverse data biases it is possible to avoid errors in extrapolating from
what we do know to inferring global patterns. Our goal here is to high-
light some promising approaches (Table 2).
Defining mass extinctions relative to the Big Five
Extinction involves both rate and magnitude, which are distinct but
intimately linked metrics
26
. Rate is essentially the number of extinctions
divided by the time over which the extinctions occurred. One can also
derive from this a proportional rate—the fraction of species that have
gone extinct per unit time. Magnitude is the percentage of species that
have gone extinct. Mass extinctions were originally diagnosed by rate:
the pace of extinction appeared to become significantly faster than
background extinction
3
. Recent studies suggest that the Devonian and
Triassic events resulted more from a decrease in origination rates than
an increase in extinction rates
4,5
. Either way, the standing crop of the
Earth’s species fell by an estimated 75% or more
2
. Thus, mass extinction,
in the conservative palaeontological sense, is when extinction rates
accelerate relative to origination rates such that over 75% of species
disappearwithin a geologicallyshort interval—typically less than 2 million
years, in some cases much less (see Table 1). Therefore, to document
where the current extinction episode lies on the mass extinction scale
defined by the Big Five requires us to know both whether current extinc-
tion rates are above background rates (and if so, how far above) and how
closely historic and projected biodiversity losses approach 75% of the
Earth’s species.
Background rate comparisons
Landmark studies
12,14–17
that highlighted a modern extinction crisis
estimated current rates of extinction to be orders of magnitude higher
than the background rate (Table 2). A useful and widely applied metric
BOX 1
Severe data comparison problems
Geography
The fossil record is very patchy, sparsest in upland environments and tropics, but modern global distributions are known for many species.
A possible comparative technique could be to examine regions or biomes where both fossil an d modern data exist—such as the near-shore marine
realm including coral reefs and terrestrial depositional lowlands (river valleys, coastlines, and lake basins). Currently available databases
6
could be
used to identify modern taxa with geographic ranges indicating low fossilization potential and then extract them from the current-extinction equation.
Taxa available for study
The fossil record usually includes only species that possess identifiable anatomical hard parts that fossilize well. Theoretically all living species
could be studied, but in practice extinction analyses often rely on the small subset of species evaluated by the IUCN. Evaluation following IUCN
procedures
34
places species in one of the following categories: extinct (EX), extinct in the wild (EW), critically endangered (CR), endangered (EN),
vulnerable (VU), near threatened (NT), least concern (LC), or data deficient (DD, information insufficient to reliably determine extinction risk).Species
in the EX and EW categories are typically counted as functionally extinct. Those in the CR plus EN plus VU categories are counted as ‘threatened’.
Assignment to CR, EN or VU is based on how high the risk of extinction is determined to be using five criteria
34
(roughly,CR probability of extinction
exceeds 0.50 in ten years or threegenerations; ENprobability of extinction exceeds 0.20 in 20 years or five generations; VU probability of extinction
exceeds 0.10 over a century
24
).
A possible comparative technique could be to use taxa best known in both fossil and modern records: near-shore marine species with shells,
lowland terrestrial vertebrates (especially mammals), and some plants. This would require improved assessments of modern bivalves and
gastropods. Statistical techniques could be used to clarify how a subsample of well-assessed taxa extrapolates to undersampled and/or poorly
assessed taxa
25
.
Taxonomy
Analyses of fossils areoften done at the level of genus rather than species. When species are identifiedthey are usually based on a morphological
species concept. This can result in lumping species together that are distinct, or, if incomplete fossil material is used, over-splitting species. For
modern taxa, analyses are usually done at the level of species, often using a phylogenetic species concept, which probably increases species
counts relative to morphospecies.
A possible comparative technique would be to aggregate modern phylogenetic species into morphospecies or genera before comparing with the
fossil record.
Assessing extinction
Fossil extinction is recorded when a taxon permanently disappears from the fossil record and underestimates the actual number of extinctions (and
numberofspecies)becausemosttaxahavenofossilrecord.Theactualtimeof extinction almost always postdates the last fossil occurrence. Modern
extinction is recorded when no further individuals of a species are sighted after appropriate efforts. In the past few decades designation as ‘extinct’
usually follows IUCN criteria, which are conservative and likely to underestimate functionally extinct species
34
. Modern extinction is also
underestimated because many species are unevaluated or undescribed.
A possiblecomparativetechniquecould be to standardize extinctioncounts by numberof species knownper time interval of interest (proportional
extinction). However, fossil data demonstrate that background rates can varywidely from one taxonto the next
35,86,87
, so fossil-to-modern extinction
rate comparisons are most reliably done on a taxon-by-taxon basis, using well-known extant clades that also have a good fossil record.
Time
In the fossil record sparse samples of species are discontinuously distributed through vast time spans, from 10
3
to 10
8
years. In modern times we
have relatively dense samples of species over very short time spans of years, decades and centuries. Holocene fossils are becoming increasingly
available and valuable in linking the present with the past
48,90
.
A possible comparative technique would be to scale proportional extinction relative to the time interval over which extinction is measured.
RESEARCH REVIEW
52|NATURE|VOL471|3MARCH2011
Macmillan Publishers Limited. All rights reserved
©2011

has been E/MSY (extinctions per million species-years, as defined in refs
15 and 27). In this approach, background rates are estimated from fossil
extinctions that took place in million-year-or-more time bins. For cur-
rent rates, the proportion of species extinct in a comparatively veryshort
time (one to a few centuries) is extrapolated to predict what the rate
would be over a million years. However, both theory and empirical data
indicate that extinction rates vary markedly depending on the length of
time over which they are measured
28,29
. Extrapolating a rate computed
over a short time, therefore, will probably yield a rate that is either much
faster or much slower than the average million-year rate, so current rates
that seem to be elevated need to be interpreted in this light.
Only recently has it become possible to do this by using palaeontology
databases
30,31
combined with lists of recently extinct species. The most
complete data set of this kind is for mammals, which verifies the efficacy
of E/MSY by setting short-interval and long-interval ratesin a comparative
context (Fig. 1). A data gap remains between about one million and about
50 thousand years because it is not yet possible to date extinctions in that
time range with adequate precision. Nevertheless, the overall pattern is as
expected: the maximum E/MSY and its variance increaseas measurement
intervals become shorter. The highest rates are rare but low rates are
common; in fact, at time intervals of less than a thousand years, the most
common E/MSY is 0. Three conclusions emerge. (1) The maximum
observed rates since a thousand years ago (E/MSY <24 in 1,000-year bins
to E/MSY <693 in 1-year bins) are clearly far above the average fossil rate
(about E/MSY <1.8), and even above those of the widely recognized late-
Pleistocene megafaunal diversity crash
32,33
(maximum E/MSY <9, red
data points in Fig. 1). (2) Recent average rates are also too high compared
to pre-anthropogenic averages: E/MSY increases to over 5 (and rises to
23) in less-than-50-year time bins. (3) In the scenario where currently
‘threatened’ species
34
would ultimately go extinct even in as much as a
thousand years, the resulting rates would far exceed any reasonable
estimation of the upper boundary for variation related to interval length.
The same applies if the extinction scenario is restricted to only ‘critically
endangered’ species
34
. This does not imply that we consider all species in
these categories to be inevitably destined for extinction—simply thatin a
worst-case scenariowhere that occurred, the extinction rate for mammals
would far exceed normal background rates. Because our computational
method maximizes the fossil background rates and minimizes the current
rates (see Fig. 1 caption),our observation that modern rates are elevatedis
likely to be particularly robust. Moreover, for reasons argued by others
27
,
the modern rates we computed probably seriously underestimate current
E/MSY values.
Another approach is simply to ask whether it is likely that extinction
rates could have been as high in many past 500-year intervals as they
have been in the most recent 500 years. Where adequate data exist, as is
the case for our mammal example, the answer is clearly no. The mean
per-million-year fossil rate for mammals we determined (Fig. 1) is about
1.8 E/MSY. To maintain that million-year average, there could be no
more than 6.3% of 500-year bins per million years (126 out of a possible
2,000) with an extinction rate as high as that observed over the past 500
years (80 extinct of 5,570 species living in 500 years). Million-year
extinction rates calculated by others, using different techniques, are
slower: 0.4 extinctions per lineage per million years (a lineage in this
context is roughly equivalent to a species)
35
. To maintain that slower
million-year average, there could be no more than 1.4% (28 intervals) of
the 500-year intervals per million years having an extinction rate as high
as the current 500-year rate. Rates computed for shorter time intervals
would be even less likely to fall within background levels, for reasons
noted by ref. 27.
Magnitude
Comparisons of percentage loss of species in historical times
6,36
to the
percentage loss that characterized each of the Big Five (Fig. 2) need to
be refinedby compensating for many differences between the modern and
the fossil records
2,37–39
. Seldom taken into account is the effect of using
different species concepts (Box 1), which potentially inflates the numbers
of modern species relative to fossil species
39,40
. A second, related caveat is
that most assessments of fossil diversity are at the level of genus, not
species
2,3,37,38,41
. Fossilspecies estimatesare frequentlyobtained by calculat-
ing the species-to-genus ratio determined for well-known groups, then
extrapolating that ratio to groups for which only genus-level counts exist.
The over-75% benchmark for mass extinction is obtained in this way
2
.
Table 2
|
Methods of comparing present and past extinctions
General method Variations and representative studies References
Compare currently measured extinction rates to
background rates assessed from fossil record
E/MSY*{7, 10, 15, 27, 62
Comparative species duration (estimates species durations to derive an
estimate of extinction rate)*{
14
Fuzzy Math*{44, 80
Interval-rate standardization (empirical derivation of relationship between
rate and interval length over which extinction is measured provides context
for interpreting short-term rates){
This paper
Use various modelling techniques, including
species-area relationships, to assess loss of species
Compare rate of expected near-term future losses to estimated background
extinction rates*{{
7, 10, 14, 15
Assess magnitude of past species losses{{42, 45
Predict magnitude of future losses. Ref. 7 explores several models and
provides a range of possible outcomes using different impact storylines{{
7, 14, 15, 27, 36, 62, 81–84
Compare currently measured extinction rates to
mass-extinction rates
Use geological data and hypothetical scenarios to bracket the range of
rates that could have produced past mass extinctions, and compare with
current extinction rates (assumes Big Five mass extinctions were sudden,
occurring within 500 years, producing a ‘worst-case scenario’ for high rates,
but with the possible exception of the Cretaceous event, it is unlikely that any
of the Big Five were this fast){
This paper
Assess extinction in context of long-term clade
dynamics
Map projected extinction trajectories onto long-term diversification/
extinction trends in well-studied clades{
This paper
Assess percentage loss of species Use IUCN lists to assess magnitude or rate of actual and potential species
losses in well-studied taxa{
This paper and refs 6, 7, 10,
14, 15, 20, 36 and 62
Use molecular phylogenies to estimate extinction rate Calculate background extinction rates from time-corrected molecular
phylogenies of extant species, and compare to modern rates
85
Fuzzy Math attempts to account for different biases in fossil and modern samples and uses empirically based fossil background extinction rates as a standard for comparison: 0.25 species per million years for
marine invertebrates, determined from the ‘kill-curve’ method
86
, and 0.21 species
35
to 0.46 species
87
per million years for North American mammals, determined from applying maximum-likelihood techniques.
The molecular phylogenies method assumes that diversification rates are constant through time and can be partitioned into originations and extinctions without evidence from the fossil record. Recent work has
demonstrated that disentanglement of diversification from extinction rates by this method is difficult, particularly in the absence of a fossil record, and that extinction rates estimated from molecular phylogenies of
extant organisms are highly unreliable when diversification rates vary among lineages through time
46,88
.
*Comparison of modern short-term rates with fossil long-term rates indicate highly elevated modern rates, but does not take into account interval-rate effect.
{Assumes that the relationship between number and kind of species lost in study area can be scaled up to make global projections.
{Assumes that conclusions from well-studied taxa illustrate general principles.
REVIEW RESEARCH
3MARCH2011|VOL471|NATURE|53
Macmillan Publishers Limited. All rights reserved
©2011

Potentially valuable comparisons of extinction magnitude could come
from assessing modern taxonomic groups that are also known from
exceptionally good fossil records.The best fossil recordsare for near-shore
marine invertebrates like gastropods, bivalves and corals, and temperate
terrestrial mammals, with good information also available for Holocene
Pacific Island birds
2,33,35,42–44
. However, better knowledge of understudied
modern taxa is critically important for developing common metrics for
modern and fossil groups. Forexample, some 49% of bivalveswent extinct
during the end-Cretaceous event
43
, but only 1% of today’s species have
even been assessed
6
, making meaningful comparison difficult. A similar
problem prevails for gastropods, exacerbated because most modern
assessments are on terrestrial species, and most fossil data come from
marine species. Given the daunting challenge of assessing extinction risk
in every living species, statistical approaches aimed at understanding what
well sampled taxa tell us about extinction risks in poorly sampled taxa are
critically important
25
.
For a very few groups, modern assessments are close to adequate.
Scleractinian corals, amphibians, birds and mammals have all known
speciesassessed
6
(Fig. 2), although species counts remain a moving target
27
.
In these groups, even though the percentage of species extinct in historic
time is low (zero to 1%), 20–43% of their species and many more of their
populations are threatened (Fig. 2). Those numbers suggest that we have
not yet seen the sixth mass extinction, but that we would jump from one-
quarter to halfway towards it if ‘threatened’ species disappear.
Given that many clades are undersampled or unevenly sampled,
magnitude estimates that rely on theoretical predictions rather than
empirical data become important. Often species-area relationships or
allied modelling techniques are used to relate species losses to habitat-
area losses (Table 2). These techniques suggest that future species
extinctions will be around 21–52%, similar to the magnitudes expressed
Cycadopsida
Mammalia
Aves
Reptilia
Amphibia
Actinopterygii
Scleractinia
Gastropoda
Bivalvia
Coniferopsida
Chondrichthyes
Decapoda
Big Five mass
extinctions
0255075100
Extinction magnitude
(percentage of species)
Figure 2
|
Extinction magnitudes of IUCN-assessed taxa
6
in comparison to
the 75% mass-extinction benchmark. Numbers next to each icon indicate
percentage of species. White icons indicate species ‘extinct’ and ‘extinct in the
wild’ over the past 500 years. Black icons add currently ‘threatened’ species to
those already ‘extinct’ or ‘extinct inthe wild’; the amphibian percentage may be as
high as 43% (ref. 19). Yellow icons indicatethe Big Five species losses: Cretaceous
1Devonian, Triassic, Ordovician and Permian (from left to right). Asterisks
indicate taxa for whichvery few species (less than 3% forgastropodsand bivalves)
have beenassessed;white arrowsshow where extinction percentages are probably
inflated (because species perceived to be in peril are often assessed first). The
number ofspecies known or assessed for each of the groups listed is: Mammalia
5,490/5,490; Aves (birds) 10,027/10,027; Reptilia 8,855/1,677; Amphibia 6,285/
6,285, Actinopterygii 24,000/5,826, Scleractinia (corals) 837/837; Gastropoda
85,000/2,319; Bivalvia30,000/310,Cycadopsida307/307; Coniferopsida 618/618;
Chondrichthyes 1,044/1,044; and Decapoda 1,867/1,867.
10,000
1,000
100
10
1
0.1
0.01
0
10710610510410310210 1
Time-interval len
g
th (years)
E/MSY
Cenozoic
fossils
CR
EN
VU
CR
EN
VU CR
EN
VU
Pleistocene
Extinctions
since 2010
Minus bats
and endemics
Figure 1
|
Relationship between extinction rates and the time interval over
which the rates were calculated, for mammals. Each small grey datum point
represents the E/MSY (extinction per million species-years) calculated from
taxon durations recorded in the Paleobiology Database
30
(million-year-or-
more time bins) or from lists of extant, recently extinct, and Pleistocene species
compiled from the literature (100,000-year-and-less time bins)
6,32,33,89–97
.More
than 4,600 data points are plotted and cluster on top of each other. Yellow
shading encompasses the ‘normal’ (non-anthropogenic) range of variance in
extinction rate that would be expected given different measurement intervals;
for more than 100,000 years, it is the same as the 95% confidence interval, but
the fading to the right indicates that the upper boundary of ‘normal’ variance
becomes uncertain at short time intervals. The short horizontal lines indicate
the empirically determined mean E/MSY for each time bin. Large coloured dots
represent the calculated extinction rates since 2010. Red, the end-Pleistocene
extinction event. Orange, documented historical extinctions averaged (from
right to left) over the last 1, 30, 50, 70, 100, 500, 1,000 and 5,000 years. Blue,
attempts to enhance comparability of modern with fossil data by adjusting for
extinctions of species with very low fossilization potential (such as those with
very small geographic ranges and bats). For these calculations, ‘extinct’ and
‘extinct in the wild’ species that had geographic ranges less than 500km
2
as
recorded by the IUCN
6
, all species restrictedto islands of less than 105 km
2
, and
bats were excluded from the counts (under-representation of bats as fossils is
indicated by their composing only about 2.5% of the fossil species count, versus
around 20% of the modern species count
30
). Brown triangles represent the
projections of rates that would result if ‘threatened’ mammals go extinct within
100, 500 or 1,000years. The lowest triangle (of each vertical set) indicates the
rate if only ‘critically endangered’ species were to go extinct (CR), the middle
triangle indicates the rate if ‘critically endangered’ 1‘endangered’ species were
to go extinct (EN), and the highest triangle indicates the rate if ‘critically
endangered’ 1‘endangered’ 1‘vulnerable’ species were to go extinct (VU). To
produce Fig. 1 we first determined the last-occurrence records of Cenozoic
mammals from the Paleobiology Database
30
, and the last occurrences of
Pleistocene and Holocene mammals from refs 6, 32, 33 and 89–97. We then
used R-scripts (written by N.M.) to compute total diversity, number of
extinctions, proportional extinction, and E/MSY (and its mean) for time-bins
of varying duration. Cenozoic time bins ranged from 25 million to a million
years. Pleistocene time bins ranged from 100,000 to 5,000 years, and Holocene
time bins from 5,000years to a year. For Cenozoic data, the mean E/MSY was
computed using the average within-bin standing diversity, which was
calculated by counting all taxa that cross each 100,000-year boundary within a
million-year bin, then averaging those boundary-crossing counts to compute
standing diversity for the entire million-year-and-over bin. For modern data,
the mean was computed using the total standing diversity in each bin (extinct
plus surviving taxa). This method may overestimate the fossil mean extinction
rate and underestimate the modernmeans, so it is a conservative comparison in
terms of assessing whethermodern means are higher. The Cenozoic data are for
North America and the Pleistoceneand Holocene data are for global extinction;
adequate global Cenozoic data are unavailable. There is no apparent reason to
suspect that the North American average would differ from the global average
at the million-year timescale.
RESEARCH REVIEW
54|NATURE|VOL471|3MARCH2011
Macmillan Publishers Limited. All rights reserved
©2011

in Fig. 2, although derived quite differently. Such models may be sensi-
tive to the particular geographic area, taxa and species-area relationship
that is employed, and have usually used only modern data. However,
fossil-to-modern comparisons using species-area methods are now
becoming possible as online palaeontological databases grow
30,31,45
.An
additional, new approach models how much extinction can be expected
under varying scenarios of human impact
7
. It suggests a broader range of
possible future extinction magnitudes than previous studies, although
all scenarios result in additional biodiversity decline in the twenty-first
century.
Combined rate–magnitude comparisons
Because rate and magnitude are so intimately linked, a critical question is
whether current rates would produce Big-Five-magnitude mass extinc-
tions in the same amount of geological time that we think most Big Five
extinctions spanned (Table 1). The answer is yes (Fig. 3). Current extinc-
tion rates for mammals, amphibians, birds, and reptiles (Fig. 3, light
yellow dots on the left), if calculated over the last 500 years (a conserva-
tively slow rate
27
) are faster than (birds, mammals, amphibians, which
have 100% of species assessed) or as fast as (reptiles, uncertain because
only 19% of species are assessed) all rates that would have produced the
Big Five extinctions over hundreds of thousands or millions of years
(Fig. 3, vertical lines).
Would rates calculated for historical and near-time prehistoric
extinctions result in Big-Five-magnitude extinction in the foreseeable
future—less than a few centuries? Again, taking the 500-year rate as a
useful basis of comparison, two different hypothetical approaches are
possible. The first assumes that the Big Five extinctions took place
suddenly and asks what rates would have produced their estimated
species losses within 500 years (Fig. 3, coloured dots on the right).
(We emphasize that this is a hypothetical scenario and that we are not
arguing that all mass extinctions were sudden.) In that scenario, the rates
for contemporary extinctions (Fig. 3, light yellow dots on the left) are
slower than the rates that would have produced each of the Big Five
extinctions in 500 years. However,rates that consider ‘threatened’ species
as inevitably extinct (Fig. 3, orange dots on the left) are almost as fast as
the 500-year Big Five rates. Therefore, at least as judged using these
vertebrate taxa, losing threatened species would signal a mass extinction
nearly on par with the Big Five.
A second hypothetical approach asks how many more years it would
take for currentextinction rates to producespecies losses equivalentto Big
Five magnitudes. The answer is that if all ‘threatened’ species became
extinct withina century, and that rate then continued unabated,terrestrial
amphibian, bird and mammal extinction would reach Big Five magni-
tudes in ,240 to 540 years (241.7 years for amphibians, 536.6 years for
birds, 334.4years for mammals). Reptiles have so few of their species
assessed that they are not included in this calculation. If extinction were
limited to ‘critically endangered’ species over the next century and those
extinction rates continued, the time until 75% of species were lost per
group would be 890 years for amphibians, 2,265 years for birds and
1,519 years for mammals. For scenarios that project extinction of
‘threatened’ or ‘critically endangered’ species over 500 years instead of
a century, mass extinction magnitudes would be reached in about 1,200
to 2,690 years for the ‘threatened’ scenario (1,209 years for amphibians,
2,683 years for birds and 1,672 years for mammals) or ,4,450 to 11,330
years for the ‘critically endangered’ scenario (4,452 years for amphi-
bians, 11,326 years for birds and 7,593 years for mammals).
This emphasizes that current extinction rates are higher than those that
caused Big Five extinctions in geological time; they could be severe enough
to carry extinction magnitudes to the Big Five benchmark in as little as t hree
centuries. It also highlights areas for much-needed future research. Among
major unknowns are (1) whether ‘critically endangered’, ‘endangered’ and
‘vulnerable’ species will go extinct, (2) whether the current rates we used in
our calculations will continue, increase or decrease; and (3) how reliably
extinction rates in well-studied taxa can be extrapolated to other kinds of
species in other places
7,20,25,34
.
The backdrop of diversity dynamics
Little explored is whether current extinction rates within a clade fall out-
side expectations when considered in the context of long-term diversity
dynamics. For example, analyses of cetacean (whales and dolphins)
extinction and origination rates illustrate that within-clade diversity has
been declining for the last 5.3 million years, and that that decline is nested
within an even longer-term decline that began some 14 million years ago.
Yet, within that context, even if ‘threatened’ genera lasted as long as
100,000years before going extinct, the clade would still experience an
extinction rate that is an order of magnitude higher than anything it has
experienced during its evolutionary history
46
.
The fossil record is also enabling us to interpret better the significance
of currently observed population distributions and declines. The use of
ancient DNA, phylochronology and simulations demonstrate that the
population structure considered ‘normal’ on the current landscape has
in fact already suffered diversity declines relative to conditions a few
thousand years ago
47,48
. Likewise, the fossil record shows that species
richness and evenness taken as ‘normal’ today are low compared to pre-
anthropogenic conditions
10,27,32,33,42,45,49
.
Selectivity
During times of normal background extinction, the taxa that suffer
extinction most frequently are characterized by small geographic ranges
and low population abundance
38
. However, during times of mass extinc-
tion, the rules of extinction selectivity can change markedly, so that
widespread, abundant taxa also go extinct
37,38
. Large-bodied animals
and those in certain phylogenetic groups can be particularly hard
hit
33,50–52
. In that context, the reduction of formerly widespread ranges
8
and disproportionate culling of certain kinds of species
50–53
may be
E/MSY
1,000,000
10
100
1
0.1
1,000
10,000
100,000
Devonian
Cretaceous
TH, CR
Triassic
Ordovician
Permian
0.01
Extinction ma
g
nitude (percenta
g
e of species)
0 20 406080100
Critically
endangered
Already
extinct Threatened
Figure 3
|
Extinction rate versus extinction magnitude. Vertical lines on the
right illustrate the range of mass extinction rates (E/MSY) that would produce
the Big Five extinction magnitudes, as bracketed by the best available data from
the geological record. The correspondingly coloured dots indicate what the
extinction rate would have been if the extinctions had happened
(hypothetically) over only 500 years. On the left, dots connected by lines
indicate the rate as computed for the past 500years for vertebrates: light yellow,
species already extinct; dark yellow, hypothetical extinction of ‘critically
endangered’ species; orange, hypothetical extinction of all ‘threatened’ species.
TH: if all ‘threatened’ species became extinct in 100 years, and that rate of
extinction remained constant, the time to 75% species loss—that is, the sixth
mass extinction—would be ,240 to 540 years for those vertebrates shown here
that have been fully assessed (all but reptiles). CR: similarly, if all ‘critically
endangered’ species became extinct in 100 years, the time to 75% species loss
would be ,890 to 2,270 years for these fully assessed terrestrial vertebrates.
REVIEW RESEARCH
3MARCH2011|VOL471|NATURE|55
Macmillan Publishers Limited. All rights reserved
©2011

particularly informative in indicating that extinction-selectivity is chan-
ging into a state characterizing mass extinctions.
Perfect storms?
Hypotheses to explain the general phenomenon of mass extinctions
have emphasized synergies between unusual events
54–57
. Common fea-
tures of the Big Five (Table 1) suggest that key synergies may involve
unusual climate dynamics, atmospheric composition and abnormally
high-intensity ecological stressors that negatively affect many different
lineages. This does not imply that random accidents like a Cretaceous
asteroid impact
58,59
would not cause devastating extinction on their own,
only that extinction magnitude would be lower if synergistic stressors
had not already ‘primed the pump’ of extinction
60
.
More rigorously formulating and testing synergy hypotheses may be
especially important in assessing sixth mass extinction potential,
because once again the global stage is set for unusual interactions.
Existing ecosystems are the legacy of a biotic turnover initiated by the
onset of glacial–interglacial cycles that began ,2.6 million years ago,
and evolved primarily in the absence of Homo sapiens. Today, rapidly
changing atmospheric conditions and warming above typical interglacial
temperatures as CO
2
levels continue to rise, habitat fragmentation, pol-
lution, overfishing and overhunting, invasive species and pathogens (like
chytrid fungus), and expanding human biomass
6,7,18,20
are all more
extreme ecological stressors than most living species have previously
experienced. Without concerted mitigation efforts, such stressors will
accelerate in the future and thus intensify extinction
7,20
, especially given
the feedbacks between individual stressors
56
.
View to the future
There is considerably more to be learned by applying new methods that
appropriatelyadjust for the different kindsof data and timescales inherent
in the fossil records versus modern records. Future work needs to: (1)
standardizerate comparisons to adjust for rate measurements over widely
disparate timescales; (2) standardize magnitude comparisonsby using the
same species (or other taxonomic rank) concepts for modern and fossil
organisms; (3) standardize taxonomic and geographic comparisons by
using modern and fossil taxa that have equal fossilization potential; (4)
assess the extinction risk of modern taxa such as bivalves and gastropods
that are extremely common in the fossil record but are at present poorly
assessed; (5) set current extinction observations in the context of long-
term clade, species-richness, and population dynamics using the fossil
record and phylogenetic techniques; (6) further explore the relationship
between extinction selectivity and extinction intensity; and (7) develop
and test modelsthat posit general conditions requiredfor mass extinction,
and how those compare with the current state of the Earth.
Our examination of existing datain these contexts raises twoimportant
points. First, the recent loss of species is dramatic and serious but does not
yet qualify as a mass extinction in the palaeontological sense of the Big
Five. In historic times we have actually lost only a few per cent of assessed
species (though we have no way of knowing how many species we have
lost that had never been described). It is encouraging that there is still
much of the world’s biodiversity left to save, but daunting that doing so
will require the reversal of many dire and escalating threats
7,20,61–63
.
The second point is particularly important. Even taking into account
the difficulties of comparing the fossil and modern records, and applying
conservative comparative methods that favour minimizing the differ-
ences between fossil and modern extinction metrics, there are clear indi-
cations that losing species now in the ‘critically endangered’ category
would propel the world to a state of mass extinction that has previously
been seen only five times in about 540 million years. Additional losses of
species in the ‘endangered’ and ‘vulnerable’ categories could accomplish
the sixth mass extinction in just a few centuries. It may be of particular
concern that this extinction trajectory would play out under conditions
that resemble the ‘perfect storm’ that coincided with past mass extinc-
tions: multiple, atypical high-intensity ecological stressors, including
rapid, unusual climate change and highly elevated atmospheric CO
2
.
The huge difference between where we are now, and where we could
easily be within a few generations, reveals the urgency of relieving the
pressures that are pushing today’s species towards extinction.
1. Novacek, M. J. (ed.) The Biodiversity Crisis: Losing What Counts (The New Press,
2001).
2. Jablonski,D. Extinctions in the fossil record. Phil. Trans. R. Soc. Lond. B 344, 11–17
(1994).
This paper summarizes, from a palaeontological perspective, the difficulties of
comparing the past and present extinctions.
3. Raup, D. M. & Sepkoski, J. J. Mass extinctions in the marine fossil record. Science
215, 1501–1503 (1982).
This is a statistical assessment of the Big Five extinction rates relative to
background rates.
4. Bambach, R. K. Phanerozoic biodiversitymass extinctions. Annu. Rev.Earth Planet.
Sci. 34, 127–155 (2006).
This paper discusses the definition of mass extinctions and mass depletions,
and the relative role of origination versus extinction rates in causing the
diversity reductions that characterize the Big Five.
5. Alroy, J. Dynamics of origination and extinction in the marine fossil record. Proc.
Natl Acad. Sci. USA 105, 11536–11542 (2008).
6. IUCN. International Union for Conservation ofNature Red List Æhttp://www.iucn.org/
about/work/programmes/species/red_list/æ(2010).
7. Pereira, H. M. et al. Scenarios for global biodiversity in the 21st century. Science
330, 1496–1501 (2010).
8. Ceballos, G. & Ehrlich, P. R. Mammal population losses and the extinction crisis.
Science 296, 904–907 (2002).
9. Hughes, J. B., Daily, G. C. & Ehrlich, P. R. Population diversity: its extent and
extinction. Science 278, 689–692 (1997).
10. Dirzo, R. & Raven, P. H. Global state of biodiversity and loss. Annu. Rev. Environ.
Resour. 28, 137–167 (2003).
This paper is an overview of the taxonomic and spatiotemporal patterns of
biodiversity and the magnitude of the current biodiversity crisis.
11. Joppa, L. N., Roberts, D. L. & Pimm, S. L. Howmany species of flowering plants are
there? Proc. R. Soc. Lond. B. doi:10.1098/rspb.2010.1004 (2010).
12. Leakey, R. & Lewin, R. The Sixth Extinction: Patterns of Life and the Future of
Humankind (Doubleday, 1992).
13. Wake, D. B. & Vredenburg, V. T. Are we in the midst of the sixth mass extinction? A
view from the world of amphibians. Proc. Natl Acad. Sci. USA 105, 11466–11473
(2008).
14. May, R. M., Lawton, J. H. & Stork,N. E. in Extinction Rates (eds Lawton,J. H. & May, R.
M. ) Ch. 1, 1–24 (Oxford University Press, 1995).
This paper compares fossil-background and recent extinction rates and
explains the numerous assumptions that are required for the comparison.
15. Pimm, S. L., Russell, G.J., Gittleman, J. L. & Brooks, T. M. The future of biodiversity.
Science 269, 347–350 (1995).
This paper explains and uses the E/MSY metric to compare the fossil-
background, current, and predicted future extinction rates.
16. Myers, N. Mass extinctions: what can the past tell us aboutthe present and future?
Palaeogeogr. Palaeoclimatol. Palaeoecol. 82, 175–185 (1990).
17. Pimm, S. L. & Brooks, T. M. in Natureand Human Society: The Quest for a Sustainable
World (eds Raven, P. H. & Williams, T.) 46–62 (National Academy Press, 1997).
18. Barnosky, A. D. Heatstroke: Nature in an Age of Global Warming 1–269 (IslandPress,
2009).
19. Vredenburg, V. T., Knapp, R. A., Tunstall, T. S. & Briggs, C. J. Dynamics of an
emerging disease drive large-scale amphibian population extinctions. Proc. Natl
Acad. Sci. USA 107, 9689–9694 (2010).
20. Hoffmann, M. et al. The impact of conservation on the status of the world’s
vertebrates. Science 330, 1503–1509 (2010).
21. Avise, J. C., Walker, D. & Johns, G. C. Speciation durations and Pleistocene effects
on vertebrate phylogeography. Proc. R. Soc. Lond. B 265, 1707–1712 (1998).
22. Weir, J. T. & Schluter, D. Thelatitudinal gradient in recentspeciation and extinction
rates of birds and mammals. Science 315, 1574–1576 (2007).
23. Lu, P. J., Yogo, M. & Marshall, C. R. Phanerozoic marine biodiversity dynamics in
light of the incompleteness of the fossil record. Proc. Natl Acad. Sci. USA 103,
2736–2739 (2006).
24. Vie
´, J.-C., Hilton-Taylor, C. & Stuart, S. N. (eds) Wildlife in a Changing World—An
Analysis of the 2008 IUCN Red List of Threatened Species 180 (IUCN, 2009).
25. Baillie, J. E. M. et al. Toward monitoring global biodiversity. Conserv.Lett. 1, 18–26
(2008).
26. S¸engo
¨r, A. M. C., Atayman, S. & O
¨zeren, S. A scale of greatness and causal
classification of mass extinctions: implications for mechanisms. Proc. Natl Acad.
Sci. USA 105, 13736–13740 (2008).
27. Pimm, S., Raven, P., Peterson, A., Sekercioglu, Ç. H. & Ehrlich, P. R. Humanimpacts
on the rates ofrecent, present, and futurebird extinctions.Proc. Natl Acad. Sci. USA
103, 10941–10946 (2006).
28. Foote, M. Temporal variation in extinction risk and temporal scaling of extinction
metrics. Paleobiology 20, 424–444 (1994).
This paper addresses the effect of interval length on extinction metrics using
simulations.
29. Foote, M. & Raup, D. M. Fossil preservation and the stratigraphic ranges of taxa.
Paleobiology 22, 121–140 (1996).
30. PBDB. The Paleobiology Database Æhttp://paleodb.org/cgi-bin/bridge.plæ(2010).
31. NEOMAP. The Neogene MammalMapping Portal Æhttp://www.ucmp.berkeley.edu/
neomap/æ(2010).
RESEARCH REVIEW
56|NATURE|VOL471|3MARCH2011
Macmillan Publishers Limited. All rights reserved
©2011

32. Barnosky, A. D. Megafauna biomass tradeoff as a driver of Quaternary and future
extinctions. Proc. Natl Acad. Sci. USA 105, 11543–11548 (2008).
33. Koch, P. L. & Barnosky, A. D. LateQuaternary extinctions:state of the debate. Annu.
Rev. Ecol. Evol. Syst. 37, 215–250 (2006).
34. Mace, G. M. et al. Quantification of extinction risk: IUCN’s system for classifying
threatened species. Conserv. Biol. 22, 1424–1442 (2008).
This paper explainsthe methodology used by the IUCN to assess the extinction
risks of extant species.
35. Alroy, J. Constant extinction, constrained diversification, and uncoordinated stasis
in North American mammals. Palaeogeogr. Palaeoclimatol. Palaeoecol. 127,
285–311 (1996).
36. Stork, N. E. Re-assessing current extinction rates. Biodivers. Conserv. 19, 357–371
(2010).
37. Jablonski, D. Lessons from the past: evolutionary impacts of mass extinctions.
Proc. Natl Acad. Sci. USA 98, 5393–5398 (2001).
38. Jablonski, D. Extinction and the spatial dynamics of biodiversity. Proc. Natl Acad.
Sci. USA 105, 11528–11535 (2008).
39. Purvis, A., Jones, K. E.& Mace, G. M. Extinction. Bioessays 22, 1123–1133 (2000).
40. Agapow, P.-M. et al. The impact of the species concept on biodiversity studies. Q.
Rev. Biol. 79, 161–179 (2004).
41. Alroy, J. et al. Phanerozoic trends in the global diversity of marine invertebrates.
Science 321, 97–100 (2008).
42. Steadman, D. W. Extinctionand Biogeography of Tropical Pacific Birds (University of
Chicago Press, 2006).
43. Raup, D. M. & Jablonski, D. Geography of end-Cretaceous marine bivalve
extinctions. Science 260, 971–973 (1993).
44. Regan, H. M., Lupia, R., Drinnan, A. N. & Burgman, M. A. The currency and tempoof
extinction. Am. Nat. 157, 1–10 (2001).
45. Carrasco, M. A., Barnosky, A. D. & Graham, R. W. Quantifying the extent of North
American mammal extinction relative to the pre-anthropogenic baseline. PLoS
ONE 4, e8331 (2009).
This paper uses species-area relationships based on fossil data to demonstrate
that the recent biodiversity baseline for mammals is substantially depressed
with respect to its normal condition.
46. Quental, T. B. & Marshall,C. R. Diversity dynamics:molecular phylogenies need the
fossil record. Trends Ecol. Evol. 25, 434–441 (2010).
47. Anderson, C. N. K., Ramakrishnan, U., Chan, Y. L. & Hadly, E. A. Serial SimCoal: a
population genetics model for data from multiple populations and points in time.
Bioinformatics 21, 1733–1734 (2005).
48. Ramakrishnan, U. & Hadly, E. A. Using phylochronology to reveal cryptic
population histories: review and synthesis of four ancient DNA studies. Mol. Ecol.
18, 1310–1330 (2009).
This paper uses a hypothesis-testing framework to reveal population histories
and compare past populations to present ones.
49. Blois, J. L., McGuire, J. L. & Hadly,E. A. Small mammal diversity loss in response to
late-Pleistocene climatic change. Nature 465, 771–774 (2010).
50. Cardillo, M. et al. Multiplecauses of high extinction risk in large mammal species.
Science 309, 1239–1241 (2005).
51. Davies, T. J. et al. Phylogenetic trees and the future of mammalian biodiversity.
Proc. Natl Acad. Sci. USA 105, 11556–11563 (2008).
52. Isaac, N. J. B., Turvey, S. T., Collen, B., Waterman, C. & Baillie, J. E. M. Mammals on
the EDGE: conservation priorities based on threat and phylogeny. PLoS ONE 2,
e296 (2007).
53. Russell, G. J., Brooks, T. M., McKinney, M. M. & Anderson, C. G.Present andfuture
taxonomic selectivity in bird and mammal extinctions. Conserv. Biol. 12,
1365–1376 (1998).
54. Erwin, D. H. The Permo-Triassic extinction. Nature 367, 231–236 (1994).
55. Arens, N. C. & West, I. D. Press-pulse: a general theory of mass extinction?
Paleobiology 34, 456–471 (2008).
56. Brook, B. W., Sodhi, N. S. & Bradshaw, C. J. A. Synergies among extinction drivers
under global change. Trends Ecol. Evol. 23, 453–460 (2008).
57. Jablonski, D. in Dynamics of Extinction (ed. Elliott, D. K.) 183–229 (Wiley, 1986).
58. Alvarez, L. W., Alvarez, W., Asaro, F. & Michel, H. V. Extraterrestrial cause for the
Cretaceous-Tertiary extinction. Science 208, 1095–1108 (1980).
59. Schulte, P. et al. The Chicxulub asteroid impact and mass extinction at the
Cretaceous-Paleogene boundary. Science 327, 1214–1218 (2010).
60. Prauss, M. L. The K/Pg boundary at Brazos-River, Texas, USA—an approach by
marine palynology. Palaeogeogr. Palaeoclimatol. Palaeoecol. 283, 195–215
(2009).
61. GBO3. Global Biodiversity Outlook 3 94 (Secretariat of the Convention on Biological
Diversity, 2010).
62. Mace, G. et al. in Millenium Ecosystem Assessment, Ecosystems and Human Well-
being: Biodiversity Synthesis (eds Ceballos, S. L. G., Orians, G. & Pacala, S.) Ch. 4,
77–122 (World Resources Institute, 2005).
63. Cardillo, M., Mace, G. M., Gittleman, J. L. & Purvis, A. Latent extinction risk and the
future battlegrounds of mammal conservation. Proc. Natl Acad. Sci. USA 103,
4157–4161 (2006).
64. Sepkoski, J. J. in Global Events and Event Stratigraphy in the Phanerozoic (ed.
Walliser, O. H.) 35–51 (Springer, 1996).
65. Sheehan, P. M. The Late Ordovician mass extinction. Annu. Rev. Earth Planet. Sci.
29, 331–364 (2001).
66. Sutcliffe, O. E., Dowdeswell, J. A., Whittington, R. J., Theron, J. N. & Craig, J.
Calibrating the Late Ordovician glaciation and mass extinction by the eccentricity
cycles of Earth’s orbit. Geology 28, 967–970 (2000).
67. Sandberg, C. A., Morrow, J. R. & Zlegler, W. in Catastrophic Events and Mass
Extinctions: Impacts and Beyond (eds Koeberl, C. & MacLeod, K. G.) 473–387
(Geological Society of America Special Paper 356, GSA, 2002).
68. McGhee, G. R. The Late Devonian Mass Extinction 1–302 (Columbia University
Press, 1996).
69. Murphy, A. E., Sageman,B. B. & Hollander, D. J. Eutrophication by decoupling of the
marine biogeochemical cycles of C, N, and P: a mechanism for the Late Devonian
mass extinction. Geology 28, 427–430 (2000).
70. Algeo, T. J., Scheckler, S. E. & Maynard, J. B. in Plants Invade the Land: Evolutionary
and Environmental Approaches (eds Gensel, P. G. & Edwards, D.) 213–236
(Columbia University Press, 2000).
71. Berner,R. A. Examinationof hypothesesfor the Permo-Triassicboundaryextinction
by carbon cycle modeling. Proc. Natl Acad. Sci. USA 99, 4172–4177 (2002).
72. Payne, J. L. et al. Calcium isotope constraints on the end-Permian mass extinction.
Proc. Natl Acad. Sci. USA 107, 8543–8548 (2010).
73. Knoll, A. H., Bambach, R. K., Payne,J. L., Pruss, S. & Fischer, W. W. Paleophysiology
and end-Permian mass extinction. Earth Planet. Sci. Lett. 256, 295–313 (2007).
74. Hesselbo, S. P., McRoberts, C. A. & Palfy, J. Triassic-Jurassic boundary events:
problems, progress, possibilities. Palaeogeogr. Palaeoclimatol. Palaeoecol. 244,
1–10 (2007).
75. Ward, P. D. et al. Sudden productivity collapse associated withthe Triassic-Jurassic
boundary mass extinction. Science 292, 1148–1151 (2001).
76. Archibald, J. D. et al. Cretaceous extinctions: multiple causes. Science 328, 973
(2010).
77. Keller, G. Cretaceous climate, volcanism, impacts, and biotic effects. Cretac. Res.
29, 754–771 (2008).
78. Mukhopadhyay,S., Farley, K. A. & Montanari, A. A short durationof the Cretaceous-
Tertiary boundary event: evidence from extraterrestrial helium-3. Science 291,
1952–1955 (2010).
79. Royer, D. L. CO
2
-forced climate thresholds during the Phanerozoic. Geochim.
Cosmochim. Acta 70, 5665–5675 (2006).
80. McCallum, M. L. Amphibian decline or extinction? Current declines dwarf
background extinction rate. J. Herpetol. 41, 483–491 (2007).
81. Rosenzweig, M.L. Loss of speciation rate willimpoverish future diversity. Proc. Natl
Acad. Sci. USA 98, 5404–5410 (2001).
82. Thomas, C. D. et al. Extinction risk from climate change. Nature 427, 145–148
(2004).
83. Pimm, S. L. & Raven, P. H. Extinction by numbers. Nature 403, 843–845 (2000).
84. Sinervo, B. et al. Erosion of lizard diversity by climate change and altered thermal
niches. Science 328, 894–899 (2010).
85. Roelants, K. et al. Global patterns of diversification in the history of modern
amphibians. Proc. Natl Acad. Sci. USA 104, 887–892 (2007).
86. Raup, D. M. A kill curve for Phanerozoic marine species. Paleobiology 17, 37–48
(1991).
87. Foote, M. Estimating taxonomic durations and preservation probability.
Paleobiology 23, 278–300 (1997).
88. Rabosky, D. L. Extinction rates should not be estimated from molecular
phylogenies. Evolution 64, 1816–1824 (2010).
89. Barnosky, A. D. & Lindsey, E. L. Timing of Quaternary megafaunal extinction in
South America in relation to human arrival and climate change. Quat. Int. 217,
10–29 (2010).
90. Turvey, S. T. Holocene Extinctions (Oxford University Press, 2009).
91. Faith, J. T. & Surovell, T. A. Synchronous extinction of North America’s Pleistocene
mammals. Proc. Natl Acad. Sci. USA 106, 20641–20645 (2009).
92. Surovell, T., Waguespack, N. & Brantingham, P. J. Global archaeological evidence
for proboscidean overkill. Proc. Natl Acad. Sci. USA 102, 6231–6236 (2005).
93. Finlayson, C. et al. Late survival of Neanderthals at the southernmost extreme of
Europe. Nature 443, 850–853 (2006).
94. Morwood, M. J. et al. Archaeology and age of a new hominin from Flores in eastern
Indonesia. Nature 431, 1087–1091 (2004).
95. Orlova, L. A., Vasil’ev,S. K., Kuz’min, Y. V. & Kosintsev,P. A. New data on the time and
place of extinction of the woolly rhinoceros Coelodonta antiquitatis Blumenbach,
1799. Dokl. Akad. Nauk 423, 133–135 (2008).
96. Reumer, J. W. F. et al. Late Pleistocene survival of the saber-toothed cat
Homotherium in Northwestern Europe. J. Vertebr. Paleontol. 23, 260–262 (2003).
97. MacPhee, R. D. E. Extinctions in Near Time: Causes, Contexts, and Consequences
(Kluwer Academic/Plenum Publishers, 1999).
Acknowledgements S. Beissinger, P. Ehrlich, E. Hadly, A. Hubbe,D. Jablonski, S. Pimm
and D. Wake provided constructive comments. Paleobiology Database data were
contributed by M. Carrano, J. Alroy, M. Uhen, R. Butler, J. Mueller, L. van den Hoek
Ostende, J. Head,E. Fara, D. Croft, W. Clyde, K. Behrensmeyer, J. Hunter,R. Whatley and
W. Kiessling. The work was funded in part by NSF grants EAR-0720387 (to A.D.B.) and
DEB-0919451 (supporting N.M.). This is University of California Museum of
Paleontology Contribution 2024.
Author Contributions All authors participated in literature review and contributed to
discussions that resulted in this paper. A.D.B. planned the project, analysed and
interpreted data, and wrote the paper. N.M. and S.T. performedkey data analyses and
interpretation relating to rate comparisons. G.O.U.W., B.S. and E.L.L. assembled critical
data. T.B.Q. and C.M. provided data, analyses and ideas relating to diversity dynamics
and rate-magnitude comparisons. J.L.M. helped produce figures and with N.M., S.T.,
G.O.U.W., B.S., T.B.Q., C.M., K.C.M., B.M. and E.A.F. contributed to finalizing the text.
Author Information Reprints and permissions information is available at
www.nature.com/reprints. The authors declare no competing financial interests.
Readers are welcome to comment on the online version of this article at
www.nature.com/nature. Correspondence should be addressed to A.D.B.
(barnosky@berkeley.edu).
REVIEW RESEARCH
3MARCH2011|VOL471|NATURE|57
Macmillan Publishers Limited. All rights reserved
©2011