Review TRENDS in Ecology and Evolution Vol.22 No.

10

Mass extinction events and the plant

fossil record
Jennifer C. McElwain1,2 and Surangi W. Punyasena3
1
School of Biology and Environmental Science, University College Dublin, Belfield, Dublin 4, Ireland
2
Department of Geology, The Field Museum, 1400 S. Lake Shore Drive, Chicago, IL 60605, USA
3
Committee on Evolutionary Biology, University of Chicago, 1025 East 57th Street, Culver 402, Chicago, IL 60637-1508, USA

Five mass extinction events have punctuated the geo-

logical record of marine invertebrate life. They are Glossary
characterized by faunal extinction rates and magnitudes
Autochthonous: formed or originating in the place where found.
that far exceed those observed elsewhere in the geo- Bennettites: a group of gymnosperms with cycad-like foliage and reproductive
logical record. Despite compelling evidence that these structures arranged into ‘flowerlike’ sometimes bisexual structures, which
were probably insect pollinated.
extinction events were probably driven by dramatic Clastic: fragments of pre-existing rocks that have been produced by the
global environmental change, they were originally processes of weathering and erosion and in general transported to a point of
thought to have little macroecological or evolutionary deposition.
Corystosperms (Corystospermales): a group of Mesozoic seed ferns that
consequence for terrestrial plants. New high-resolution were most abundant in the Triassic. The foliage – Dicroidium, Pachypteris
regional palaeoecological studies are beginning to chal- and Stenopteris– is characterized by pinnately divided leaves with open
lenge this orthodoxy, providing evidence for extensive venation and both pollen (Pteruchus) and ovulate (Umkomasia) organs are
branched.
ecological upheaval, high species-level turnover and Czekanowskia: an extinct Mesozoic genus characterized by highly dissected
recovery intervals lasting millions of years. The chal- (filiform) leaves containing a single vein in each dichotomy. Leaves are borne
lenge ahead is to establish the geographical extent of on short shoots, and ovules structures (Leptostrobus) are stalked and protected
within a so-called capsule. The habit of Czekanowskia is unknown but generally
the ecological upheaval, because reconstructing the presumed to be woody.
vegetation dynamics associated with these events will Glossopterids (Glossopteridales): an extinct group of predominantly Permian
elucidate the role of floral change in faunal mass extinc- seed ferns with tongue-shaped leaves with entire laminas, prominent midrib
and reticulate venation. Leaves were borne in whorls, probably on short shoots
tion and provide a better understanding of how plants and the predominant plant habit was woody. Ovulate and pollen organs are
have historically responded to global environmental borne on or in the axil of the leaves.
Lycopsids (club mosses): also known as Lycophytes. Vascular land plants that
change similar to that anticipated for our future.
reproduce by spores. Their life cycle consists of free-living sporophyte and
gametophyte phases. Free water is required to carry out fertilization of the egg
Introduction by flagellate sperm in the gametophyte phase of the life cycle. Modern
Global marine biodiversity has been devastated five times representatives of this group are exclusively herbaceous; however, fossil
representative included herbaceous and woody habits.
during Earth history, first at the end-Ordovician (446 Mya) Mass extinction: intervals in the geological record when the relative decline in
and then again at the Frasnian–Famennian (FFB, global diversity, as measured by the percentage of global taxonomic
371 Mya), the Permian–Triassic (PTB, 251 Mya), the extinction, far exceeds the percentage of extinction typically observed during
all other intervals [86]. Typical or normal levels of extinction are often referred
Triassic–Jurassic (TJB, 200 Mya) and Cretaceous–Paleo- to as ’background‘ extinction. It remains debatable whether mass extinctions
gene boundaries (KPB, 65 Mya). These events, termed are in a statistically separate class from ‘background’ extinctions or represent
the uppermost tail of a continuous distribution of extinction intensities.
‘mass extinctions’, reset the evolutionary trajectories of
Furthermore, it has been suggested that only three of the so-called ’big five‘
marine families and orders and resulted in the collapse mass extinction events were generated by increased extinction alone (end-
and reordering of entire marine ecosystems [1]. They are Ordovician, PTB, KPB) and that TJB and FFB biodiversity loss was caused more
by low origination than by increased extinction [87].
characterized in the palaeontological record by the global Peltasperms (Peltaspermales): an extinct group of Triassic seed ferns, with
extinction of higher taxonomic groups at rates and magni- fern-like leaves – known as Lepidopteris – that reproduce by seeds. They are
tudes greatly exceeding background levels [2]. Although characterized by blisterlike swellings on both vegetative and reproductive
structures. The ovules are borne on stalked circular disks and leaves are
there is currently no clear consensus on the precipitating bipinate with open venation.
causes of these ‘big five’ mass extinction events, there is Sphenobaiera: an extinct genus belonging to the Ginkgo family which was
growing acceptance that they were probably forced by abundant and speciose in the Mesozoic, with leaves characterized by a short
pseudopetiole and a long leaf lamina typically divided into four or more coarse
abiotic change of a magnitude similar to the greenhouse- dissections with multiple veins. Leaves of some species are borne on short
induced climate change anticipated in the near future [3] shoots, as in modern Ginkgo biloba.
(Table 1), making them a model for understanding biotic Stages: refers to a package of rocks defined by a particular faunal group with a
duration of millions of years.
response to all forms of extreme environmental change. Taphonomic bias: the degree to which the factors involved in the transforma-
Among terrestrial faunas, mass extinctions have closely tion of a living organ or organism or community to a fossil organ or organism
paralleled those of marine biota, occurring relatively syn- or community affect its representation in the fossil record.
Volzialean conifers (Volziales): a group of extinct Palaeozoic to Mesozoic
chronously and with similar magnitudes of biodiversity gymnosperms with simple or forked needle-like leaves and reproductive
organs organized as ovulate or pollen cones. The group radiated in the
Permian, and the modern Norfolk Island pine provides a good analogue for the
Corresponding author: McElwain, J.C. ([email protected]). probable habit for some members of this group.
Available online 4 October 2007.

www.sciencedirect.com 0169-5347/$ – see front matter ß 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.tree.2007.09.003
 Review TRENDS in Ecology and Evolution Vol.22 No.10 549

Table 1. Comparison of percentage extinctions between plants and animals

Magnitude of global faunal extinction Magnitude of floral extinction
(family)
Mass extinction event Terrestrial fauna a Marine fauna,b Global (family) c Regional (family and/or genera)
Ordovician n.a. 22–24% n.a. n.a.
FFB 44% 21–28% Negligible 30–40%* d
PTB 62% 47–50% Negligible 19%** e
TJB 22% 20–23% Negligible 17%** f
KPB 6% 14–15% Negligible 18–30%* g
n.a. not applicable; *data based on pollen and spore record; **data based on fossil leaf record; abased on ref. [4]; bbased on ref. [2]; cbased on ref. [5]; dbased on ref. [90]; ebased
on ref. [39]; fbased on ref. [13]; gbased on ref. [15].

loss [4] (Table 1). By marked contrast, global taxonomic mass extinction boundaries, despite growing indications of
losses above the rank of genus are relatively rare among concurrent ecological upheaval in terrestrial plant com-
plants across faunal mass extinction boundaries [5–7] munities. Eighteen years since Traverse’s initial obser-
(Table 1). Prior reviews of the fossil plant record have vation [9] we ask, do plants really ‘dance to a different
identified numerous reproductive, physiological and beha- beat’ during these unique intervals in Earth history? We
vioural traits unique to plants that enable small popu- compare the evolutionary and macroecological dynamics of
lations, and hence plant species, to persist in the face of terrestrial vegetation associated with extreme environ-
extreme environmental changes [6–9]. mental and global climatic change spanning three of the
Although the loss of plant diversity is less than that big five marine mass extinction events (the PTB, TJB and
observed in the faunal record, recent palaeoecological stu- KPB), for which detailed plant palaeoecological studies are
dies of local and regional vegetation dynamics across three of available. These extinction boundaries share several abio-
the ‘big five’ marine mass extinction events, at the PTB [10– tic attributes, which were probably caused by intense
12], TJB [13] and KPB [14–17], strongly challenge the long- episodes of flood basalt volcanism, or a meteorite impact
held view that terrestrial vegetation remained unscathed in the case of the KPB (Table 2). The end-Ordovician and
after these events. These new studies, focusing on changes FFB extinctions are excluded, as vascular land plants had
in abundance and community structure in addition to taxo- not yet evolved in the case of the former and no quanti-
nomic richness, indicate ecological instability in plant com- tative plant palaeoecological data are available in the case
munities and collapse, before or coincident with peak faunal of the latter. We also do not focus on time intervals for
extinction at the PTB [10–12], TJB [13–18] and KPB [14– which there is good evidence for increased plant extinc-
17]. The results are forcing us to question whether global tions but an absence of faunal mass extinction, such as the
compilations of plant diversity [5–19], now over 20 years old, Westphalian–Stephanian boundary (305 Mya) [20,21],
are the appropriate means of understanding past vegetation early Oligocene (29 Mya, Niklas,1997) and many others.
changes. Taxonomic diversity, often used as a metric of An excellent summary of these noncoincident plant extinc-
ecosystem health for modern communities, may not be tions is available in [6]. The primary focus of our review is
the best measure of ecological change or response in the on contrasting macroevolutionary and macroecological
fossil record. Furthermore, the survival of a plant clade patterns between flora and fauna in the context of the
across a mass extinction event does not ensure provision extreme environmental change associated with the PTB,
of the same quantity or quality of primary productivity if a TJB and KPB.
once ecologically dominant clade becomes rare. In this
regard, regional palaeoecological records incorporating Vegetation dynamics across marine mass extinction
changes in abundance structure of a community may pro- boundaries
vide a more informative method of assessing past vegetation In the following section we review a suite of recent studies
change and feedback effects on faunal extinctions. documenting the taxonomic composition, palaeoecology
Here, we explore the apparently contradictory evidence and structure of ancient plant communities (based predo-
for the remarkable persistence of higher plant taxa across minantly on fossil leaves, pollen and spores) representing

Table 2. Environmental characteristics of PTB, TJB and KPB mass extinction events
Mass Average carbon isotopic shift Change in atmospheric CO2 Meteorite impact Volcanism Global temperature change
extinction in organic and inorganic C concentration (ppmV) involved in involved in
event extinction? a extinction? b
PTB 3–5% negative c >2000 ppmV increase Equivocal Yes >+ 8 8C e
across boundary d Siberian traps
h
TJB Double negative excursion of 700–1300 ppmV increase Equivocal Yes >+ 5 8C
2–3% coincident with the across boundary g Central Atlantic
boundary f Magmatic Province
KPB 1.45–2% negative excursion 700 ppmV increase 0.5 Strong evidence Yes Warming between
coincident with KPB i Mya before boundary with Deccan traps 65.9 and 66 Mya
little or no change across followed by cooling
boundary j immediately before
and across KPB k
a
based on ref. [51]; bbased on ref. [79]; cbased on refs [71–92]; dbased on refs [93–95]; ebased on refs [69–93], fbased on refs [37–96]; gbased on ref [37]; hbased on refs [37–97];
i
based on refs [73–98]; jbased on refs [99]; kbased on refs [100].

www.sciencedirect.com
550 Review TRENDS in Ecology and Evolution Vol.22 No.10

time intervals before, during and after the three most and not necessarily to a true catastrophic decline in global
recent marine mass extinction events in Earth history, plant richness [25]. These results alone would confirm the
at the PTB, TJB and KPB. established view of the persistence of higher plant clades
across extinction boundaries. However, plant response to
Permian–Triassic (251 Mya) the rapid (<500 kyrs) environmental perturbations of the
Extinctions among marine fauna at the PTB resulted in PTB [6] (Box 1, Table 2) are unlikely to be reflected in such
the most profound reorganization of marine communities broad time bins (spanning 5–6 million years) [6].
since the end-Ordovician [22] and was marked by a There are, however, additional studies on terrestrial
dramatic disturbance in the global biogeochemical cycle plant responses at the PTB at sub-million-year timescales
of carbon between rock, atmosphere and biosphere reser- that are beginning significantly to advance our under-
voirs (Box 1). This is clearly reflected by a 9% (per mil) standing of the fine-scale dynamics of terrestrial ecosys-
negative carbon isotopic excursion in the organic and tems in response to global warming and deteriorating
inorganic carbon rock records [23] (Box 1, Figure 1). environmental conditions (Box 1, Table 2). High-resolution
Approximately 50% of marine invertebrate families studies of pollen and spores from Greenland document a
became extinct at the boundary [2]. Comparable losses pulsed die-back of the dominant Permian gymnosperm
occurred on land, where between 60% and 70% of terres- vegetation before and coincident with the large negative
trial vertebrate families were lost [4–22]. The PTB also carbon isotopic excursion found at the PTB (Figure 1) [11].
marks the most significant decline in insect diversity The palynological record suggests that wooded terrestrial
experienced in the evolutionary history of that class [24]. ecosystems took four to five million years to reform stable
Global compilations of terrestrial plant family diversity ecosystems [10], while spore-producing lycopsids (Pleuro-
within geological stages preceding and following the PTB meia), ancient relatives of the modern-day quillwort Iso-
reveal a pattern consistent with that seen in the faunal etes, had an important ecological role in the post-extinction
record. Plant species richness declines more than 50% [25]. interval [10–27]. Pleuromeia had multiple morphological
However, most of this diversity loss can be accounted for by adaptations that indicate stress tolerance, including the
the loss of fossiliferous localities due to changes in palaeo- ability to take in carbon through its root system [10].
latitude, fossil preservation, and tectonic plate movements An unusually high abundance of lycopod spores that
were not dispersed individually (the normal condition) at
Box 1. Environmental changes associated with mass
the PTB but instead remained in undispersed units of four,
extinction events
known as spore tetrads [28,29], has led to the hypothesis
that meiotic processes in spore development required to
Determining the causal mechanism or mechanisms of mass produce four separate spores were disrupted at the PTB,
extinction remains an elusive goal. Meteorite impacts, catastrophic
perhaps owing to increased UV-B radiation [29]. The
methane release from marine sediments, volcanism, sea-level
change, oceanic anoxia and global wildfire have all been proposed morphological changes involved in forming such tetrads
as explanations, both alone and in combination (Table 2, main text). are minimal [28]. Abnormal gymnosperm pollen grains
Despite years of investigation, there is little consensus (with the recorded at the close of the Permian in both China and
notable exception of a meteorite impact at the KPB) on the Russia provide corroborating evidence of deterioration in
geological or extraterrestrial factor or combination that acted as a
proximal kill mechanism for faunal extinctions. However, this
global atmospheric conditions at this time [30].
pursuit of a mechanistic understanding of mass extinctions has The proliferation of Reduviasporonites spores at the
produced a wealth of high-resolution environmental records that PTB in Europe [31,32] and South Africa [33] was initially
provide a firm abiotic framework within which to examine the interpreted as a spike in saprotrophic fungal activity after
evolutionary and ecological responses of terrestrial plants across the loss of woody forests [31]. Together with evidence for
mass extinction events.
The four mass extinction events of relevance to vascular plant
pocket-rot cavities in fossil wood [26] and a loss of once-
evolution at the FFB, PTB, TJB and KPB are all characterized by large dominant woody conifer pollen [11], this fungal prolifer-
excursions (departure from normal) in stable carbon isotopic ation seemed to support the scenario of extensive forest
composition (PTB, TJB and KPB illustrated in Figure 1, main text), loss at the extinction boundary [31]. This interpretation
indicating intense disturbance of the global carbon cycle between
has been called into doubt given recent morphological and
the atmosphere, biosphere and rock reservoirs (Table 2, main text).
Dramatic changes in global climate, atmospheric composition geochemical analyses indicating that Reduviasporonites is
(carbon dioxide and oxygen), air quality (owing to acid rain, more likely to be of algal rather than of fungal origin [34]. A
rampant wildfires and high sulfate aerosols and particulate loading) worldwide algal pulse suggests widespread PTB anoxia
and light quality (e.g. global dimming owing to high particulate and/ and generally stressed environments rather than
or dust content, increased ultraviolet-B radiation owing to ozone
increased fungal breakdown of woody tissues. However,
depletion by volcanically released hydrothermal organohalogens)
have also been invoked [29–79] using geochemical and proxy cases of extensive (12 000 km2) regional die-off of modern
environmental and biological data (Table 2, main text). Many of forest dominants such as pinyon pine due to environmental
these proposed environmental changes are the expected conse- stress from drought and bark beetle infestation [35]
quences of either flood basalt volcanism [29–79], a meteorite impact indicate that the massive loss of PTB forests may not be
[51] or both, which is why it has been difficult to decipher ultimate
causal mechanisms of mass extinction. The PTB, TJB and KPB
an unrealistic scenario (Table 2).
coincide temporally with intense episodes of flood basalt volcanism Notably, instability in terrestrial plant communities,
[79], when hundreds of cubic kilometers of basalt lava were released reflected in the widely fluctuating dominance relationships
over large portions of the Earth’s surface. Meteorite impacts have among plant taxa, began millions of years before woody
been proposed for all three, although evidence for impact is Permian species of cordaites, pteridosperms and conifers
equivocal for all but the KPB [51] (Table 2, main text).
finally became extinct [11,12]. These marked shifts in plant
www.sciencedirect.com
 Review TRENDS in Ecology and Evolution Vol.22 No.10 551

Figure 1. Changes in dominance structure of regional plant communities (a–q) versus changes in carbon isotope values (d13Corg, vertical red lines) across the PTB, TJB and
KPB; only x-axis values are to scale. The relative timing of floral and faunal extinction (indicated by grey horizontal bar) and peak negative d13Corg values are approximated
to illustrate similarities and differences among the three mass extinction events. Fluctuations in d13Corg reflect changes in the global carbon cycle (Box 1). Ecologically
dominant components of the plant communities before and after PTB, TJB and KPB are illustrated using stylized illustrations of fossil plants. The order of appearance of
taxa (a to q) indicates the relative abundance of the taxon within vegetation from highest (left) to lower (right). The taxa are; (a) conifer; (b) pteridosperm; (c) Cycadales;
(d) Pleuromeia; (e) pteridophyte; (f) cordaites; (g) Sphenobaiera; (h) Podozamites; (i) Bennettitales (macrophyllous); (j) Baiera; (k) Hamamelidae; (l) Cornales; (m) monocot
(herbaceous); (n) fungi; (o) magnolid flower; (p) megaphyllous angiosperm leaves; (q) evergreen angiosperm. Redrawn with permission from [71–88,89]. ‘Disaster taxa’ are
long-ranging opportunistic taxa that become unusually abundant during recovery intervals after mass extinction because of relaxation of competition.

dominance patterns recorded in the pollen and spores Australian localities indicate extinction of dominant glos-
appear to have had a more important effect on ecosystem sopterid and cordaite vegetation and their replacement by
functioning, such as biogeochemical cycles (Box 3), than did shrubby pteridosperms (including peltasperms and corys-
the final disappearance of these once-dominant Permian tosperms), volzialean conifers and lycopsids [26]. The gen-
plant groups. This finding suggests that the abundance eral paucity of post-boundary plant palaeoecological
and dominance structure of plant communities, rather studies, however, makes uncertain the extent to which
than total biodiversity, was the most important criterion these vegetation changes are the product of migration or
for measuring ecosystem health. local facies change rather than a global pattern of veg-
Parallel changes in vegetation are observed in the etation response.
southern hemisphere, although the lack of quantitative One of the most remarkable observations for this inter-
analyses of vegetation change makes direct comparisons val is the complete global absence of coals during the early
difficult. In parts of Gondwana (now modern Antarctica), Triassic, termed the ‘coal gap’ [39,40]. The absence of peat
the dominant broad-leafed deciduous glossopterid forests formation (and hence coal in the rock record) has been
were replaced by needle-leafed conifers (Volziopsis) and attributed to the extinction of peat-forming plants such as
later by a more distantly related pteridosperm genus, glossopterids [39]. It is also possible that this gap reflects a
Dicroidium, with highly dissected leaves [36]. The change change in one of the abiotic factors required for peat
in overall leaf shape may reflect adaptation to atmospheric deposition, such as increased input of clastic sediment
acidification [36], or could also reflect adaptation to glob- from higher erosion or decreased precipitation rates. The
ally warmer climates. Smaller or more dissected leaf sizes reasons for the coal gap thus remain largely unresolved.
are an effective way of reducing the likelihood of reaching Notably, the cessation of coal deposits in the latest Per-
lethal leaf temperatures in a world with high greenhouse mian of Australia and Antarctica precedes the globally
gas concentration [37]. In Australia, a 97% species-level recognized negative carbon isotopic excursion [26–39],
extinction is recorded in fossil leaf data, although fossil suggesting, as was the case in the northern hemisphere,
pollen samples reveal a more moderate 19% taxonomic that terrestrial plant communities underwent changes
loss at genus and/or family levels [38]. As in Antarctica, in vegetation composition before peak faunal extinction
www.sciencedirect.com
552 Review TRENDS in Ecology and Evolution Vol.22 No.10

rates and minimum carbon isotopic values had been

reached.

Triassic–Jurassic (200 Mya)

The third greatest mass extinction event of marine invert-
ebrates occurred at the TJB, marked by a global 2–3%
negative carbon isotopic excursion [41] (Box 1, Figure 1,
Table 2) and the loss of 23% of marine and terrestrial
faunal families [42]. By contrast, only one plant family
(Peltaspermaceae) became globally extinct at the TJB [13–
43], and even this is questionable based on the recent
phylogenic placement of the peltasperms within the cor-
ystosperms, a clade of seed ferns that survives the TJB
[44].
The TJB is the least well studied of the three faunal
mass extinction events discussed here and is particularly
lacking with regard to plant palaeoecological records. How-
ever, available information hints at a major palaeoecolo-
gical disturbance of plant communities. Studies of plant
taxonomic ranges based on fossil pollen, spores [18] and
leaves [37–43] have revealed that up to 90% of species
became locally to regionally extinct across North America
[18–45], Greenland [43–46] and Europe [47]. Available
studies reveal parallels with the PTB. In Greenland, con-
siderable changes in the dominance and diversity struc-
ture of Triassic forests have been identified before the TJB
[13]. Four separate basins in North America [18] show a
similar pattern, where high diversity pollen assemblages
of monosulcates and monosaccates give way to lower diver-
sity assemblages dominated by Classopollis, a pollen type
normally associated with hot and/or arid climate con-
ditions [48,49]. These low diversity pollen assemblages
are confined to within twenty-one thousand years before
the TJB as defined by a fern spike [18–50] and negative
isotopic excursion [45].
Forests in East Greenland underwent a complete turn-
over in dominance across the TJB [13]. High diversity
Triassic communities of Podozamites, a broad-leafed con-
ifer, and bennettites (Pterophyllum and Anomozamites)
(Figure 2) were replaced by lower diversity and less even Figure 2. Reconstruction of (a) late Triassic (205 Mya) and (b) late Cretaceous
(i.e. more dominant) forests, overwhelmingly dominated by (66 Mya) vegetation of East Greenland before the TJB extinction and Denver before
the KPB extinction events, respectively. Illustrated by scientific illustrators Marlene
taxa that had been relatively minor components of late
Hill Donnelly (a) and Gary Staab (b). Reproduced with permission from M.H.
Triassic forests: Czekanowskia, Sphenobaiera and an Donnelly (a) and K. Johnson (DMNS) (b). Taxa illustrated are; (c) Podozamites
osmundaceous fern (Todites). Whether this signature of sp.(broad-leafed conifer), (d) Nilssonia sp. (cycad), (e) Stachyotaxus
septentrionalis (conifer), (f) Pseudoctenis spectabilis (cycad), (g) Dictyophyllum
ecological upheaval at a single locality in East Greenland
exile (dipteridaceous fern), (h) Pterophyllum sp. (bennettite), (i) herbaceous
scales up to a regional or continental level remains an monocot (angiosperm), (j) ‘Zizyphus’ fibrillosus (woody angiosperm),
important area of future research. The influence of differ- (k) Allantodiopsis erosa (climbing fern), (l) herbaceous angiosperm.
ences in the depositional setting on pre- and post-TJB
plant communities also remains to be further explored
(Box 2). regional extirpation, in other words, migration out of the
immediate area, or true global extinction. Irrespective of
Cretaceous –Paleogene (65 Mya) this uncertainty, the magnitude and pace of species-level
The KPB is most famous for the extinction of non-avian turnover in KPB fossil floras across North America is
dinosaurs and for the meteorite impact that precipitated it consistent with sudden ecosystem collapse. Plant extirpa-
[51]. However, marine communities were also devastated, tions or extinction rates at the KPB are orders of magni-
with the loss of up to 15% of marine families [2–4]. tude greater than background Cretaceous extinction rates
Although, vegetation changes at the KTB are the most [15–52]. Palaeobotanical studies of fossil pollen, spores and
intensively studied of all mass extinction boundaries, the leaves from North American localities have shown that
extent of spatial information is limited by the availability 18–30% of plant genera and families [15–52] and up
of fossiliferous rock strata. Hence, it is not always possible to 57% of plant species disappeared at the KPB [14].
to discern whether basin-level plant extinctions represent Loss of ecological dominants within floras [53–56], loss
www.sciencedirect.com
 Review TRENDS in Ecology and Evolution Vol.22 No.10 553

(based on ecological rather than diversity measures of

Box 2. Filters of fossilization
recovery) [10]. Despite the initial massive loss of species
One of the biggest obstacles to improving our understanding of diversity, few to no flowering plant families were lost and
vegetation dynamics spanning mass extinction events is the nature of species diversity did eventually recover [64], albeit dia-
the fossil record. The ecological information contained within the
fossil record is distorted by the numerous biases introduced by
chronously across North America. Johnson [54] has
transport, differential decay, probability of burial and fossilization hypothesized that wetland habitats may have served as
potential [80,81]. A plant assemblage will overrepresent robust refugias and centres of radiation in the post-extinction
specimens and material that is easily transportable [80–82]. Whole interval, as many mire-adapted taxa survived the initial
vegetation types, such as high elevation and desert biomes, are environmental changes of the KPB. Post-KPB survivors
almost completely absent from the fossil record, owing to a dearth of
suitable depositional settings. There are, however, inherent advan-
were minor elements of Cretaceous floras [14] including
tages to the plant record that are not shared with most other sources mainly Cornales, Hamamelidae and palms in the Denver
of terrestrial data, including vertebrate specimens. Leaf beds are often region and herbaceous monocots and ferns in New Mexico
preserved in situ with little transport or time-averaging [80,81], [58].
providing the closest snapshots available of terrestrial biota. Fossil
pollen and spore data provide one of the only sources of continuous
terrestrial information, including quantitative information on changes Looking for common patterns of vegetation responses
in relative abundance, despite limitations in taxonomic resolution. In Terrestrial systems experienced destabilization and col-
combination, these independent sources of data can provide a lapse at the PTB, TJB and KPB. A common macroecolo-
comprehensive view of past floras when interpreted within the gical pattern was the permanent or temporary loss of
context of taphonomic and depositional information.
ecological dominants from terrestrial plant communities
Studies have demonstrated that differences in depositional
environment account for only 15% of the observed species-level at local to regional spatial scales, which took place over
change at the KTB [15]. At the PTB the most detailed records of thousands to millions of years. The recovery intervals after
ecological upheaval in the terrestrial vegetation are derived from all three reviewed mass extinction boundaries also high-
pollen and spores, a record that is considerably less susceptible to light important similarities but also many unknowns.
the vagaries of depositional environment than are fossil leaves.
Preliminary studies of the influence of taphonomic filters on the
Although species diversity rebounded relatively rapidly
interpretation of TJB vegetation dynamics suggest that pre- and after PTB and TJB when compared with the KPB, the
post-boundary differences in depositional environments cannot time interval for full ecological recovery was on the order of
account for the observed compositional shift in the vegetation millions of years, and the TJB and PTB show some signs of
[13]. Tracking vegetation dynamics in so-called ‘isotaphonomic’ [70] prolonged ecosystem instability after the extinction
fossil plant assemblages (that is from the same depositional
environments) is an ideal future goal. However, realistically, for
horizon. In Greenland, post-TJB ecosystem instability is
mass extinction boundaries, it is unlikely that this will be possible reflected by an order of magnitude increase in the pro-
except in the rarest of localities. portion of singletons (i.e. species only occurring in one fossil
plant bed) in each time slice relative to total regional plant
diversity sampled for that same time slice [13]. This finding
of specialized insect taxa [57], catastrophically rapid loss of suggests that plant community composition was in a con-
plant species richness [14,15] and major compositional siderable state of species flux following the TJB [13]. A
changes in plant communities [58] occurring contempor- rapid diversity rebound after the PTB has been attributed
aneously in both northern and southern hemispheres all to an influx of weedy generalist ‘disaster taxa’, which were
point to global palaeoecological upheaval and rapid eco- eventually replaced by a woody conifer-dominated veg-
system failure at the KPB [14–56,58] (Figure 1). etation, similar in structure to pre-PTB conifer dominated
Extinction at the KPB appears to have been highly se- forests, four–five million years later [10–12].
lective. Extinction risk may have been higher among insect- The homogeneous low-diversity vegetation that charac-
pollinated than among wind-pollinated members of import- terized most Paleocene localities better fits expectations of
ant Cretaceous plant clades [15–59]. Studies of leaf cuticle an impoverished post-extinction ‘recovery-phase’ veg-
and leaf morphospecies indicate that predominantly ever- etation. However, the recent discovery of an exceptionally
green angiosperm families, including Laurales, Illiciales rich Paleocene aged fossil flora with tropical-rainforest-like
and Cyclanthales, experienced the highest levels of species leaves at Castle Rock in Denver, just one million years
extinction; surviving species within these and other families after the KTB challenges this view [64]. It also emphasizes
were mainly deciduous [58–61]. Peak abundance of fungal the possible pitfalls of extrapolating from a single locality
spores coincident with the iridium anomaly in North Amer- to regional and global scales, and highlights the import-
ica [62] and New Zealand [16] suggest that large-scale ance of increasing spatial sampling of fossil floras spanning
regional and perhaps even global dieback of terrestrial all mass extinction boundaries. How such high diversity
vegetation occurred at the KPB. The ‘fungal spike’ is quickly was attained in the Castle Rock flora so soon after the
followed by a so-called ‘fern spike’ [14–62] or high abun- apparently complete ecosystem collapse at the KTB
dances of weedy generalists, often herbaceous vegetation remains an important but unresolved issue.
[63], which are interpreted as evidence for recolonization of Although few plant families are lost across all three
the landscape by ferns, herbaceous angiosperms, or spore- boundaries, extinction at the species level was highly se-
producing aquatics. lective. Several characteristics are shared among the most
Post-boundary fossil floras were universally species vulnerable taxa. Reproductive specialization was one
poor for over one million years across the US western important factor, as evidenced by survival of wind-polli-
interior [14], suggesting a long recovery interval, although nated taxa over insect-pollinated taxa at the KPB [17–59]
one that is shorter than that observed following the PTB and the ecological demise of the most reproductively
www.sciencedirect.com
554 Review TRENDS in Ecology and Evolution Vol.22 No.10

specialized Triassic taxa, including a majority of probably but cannot as yet be readily explained. For instance,
insect-pollinated Bennettitales and Cycadales at the TJB ecological rarity (defined as having relative abundances
[13]. Woody pre-boundary vegetation was temporarily in the last quartile of rank abundance distributions [65]), a
replaced by herbaceous and/or pteridophyte (spore-produ- trait predicted by modern metapopulation models to
cing) communities after PTB and KPB. The most consist- increase the risk of extinction, appears to have heightened
ent pattern to emerge is that most post-boundary survivors proneness to extinction at the TJB [13] but not at the PTB
were of low abundance before the extinction boundary, [10]. The case for the KPB is somewhat equivocal [15],
indicating that they were probably poor competitors in because although rare Cretaceous gymnosperms, in-
pre-boundary time, but had a significant ecological role cluding Elatides, Dammarites and Nilssonia, disappeared
in the post-boundary vegetation at all three mass extinc- at the boundary in North America [53], low abundance is
tions (KPB [53]; TJB [13]; PTB [11]). This pattern suggests not a good predictor of extinction risk among angiosperm
that they were tolerant of environmental stress and/or clades [15].
released from competition by Triassic dominants. Plant families and orders demonstrate incredible resili-
Striking macroecological differences in the vegetation ence across the PTB, TJB and KPB mass extinction events
dynamics spanning all three boundaries are also apparent (Table 1) and, from this respect, do not therefore undergo
‘mass extinction’ in the strictest definition of the term.
However, available plant palaeoecological data document
Box 3. Did vegetation change contribute to faunal mass large-scale and profound ecological change associated with
extinction? all three extinction boundaries. These ecological changes
A recent suite of sedimentological studies have suggested that there were probably of sufficient magnitude to have significantly
is evidence for an abrupt change in the style of river drainage across impacted ecosystem function at the local to regional level
the Permian–Triassic boundary in numerous Southern hemisphere and may well have contributed to losses in terrestrial
basins [36–83]. A switch from low sinuosity meandering to braided
river systems has been largely attributed to extinction-driven
faunal diversity by altering primary productivity [66]
vegetation changes, which increased sediment flux into the rivers, and marine faunal diversity by feedback effects on run-
causing the change in fluvial style. An alternative explanation of the off rates and biogeochemical cycles. Evidence for disrupted
observed pattern is that dramatic compositional shifts among land biogeochemical cycles, which hint at altered functioning of
plants and loss of taxonomic diversity across the Permian–Triassic
the terrestrial ecosystem, include dramatic changes in
boundary is a reflection of the fossil sample availability that would
be expected to decrease with the change in fluvial style. We are river drainage patterns in PTB sections in Antarctica,
therefore left with a chicken and egg debate – did the vegetation Australia and South Africa [36–69] (but see [70]), and
changes result in altered river morphology and drainage style or did increased particulate [71] and nutrient flux [23] from land
the fluvial changes alter the predominant depositional environment into the marine system at the PTB and FFB boundary,
which, in turn, affected how representative the fossil plant
respectively (see Box 3 for more details). Exploring lin-
assemblages before and after the boundary are of the true
dominance diversity structure of the source vegetation? It is likely kages between plant ecological change, biogeochemical
that the real answer is more nuanced than either camp would be cycles and marine mass extinction clearly warrants further
willing to acknowledge. In the case of the PTB, preliminary evidence investigation, because they are relevant to our understand-
supporting the hypothesis that changes in terrestrial ecosystems ing of modern systems.
contributed to marine faunal extinction comes from fossil evidence
for major coastal cyanobacterial blooms, which might indicate an
increased influx of terrestrially derived nutrients, although they Conclusions
might also be a result of decreased grazing pressure by marine In the marine realm, mass extinctions removed the domi-
invertebrates [23]. An unusual pulse of terrestrial polysaccharides nant incumbent genera, families and even orders via
into the marine system, coincident with the negative carbon isotopic extinction [72]. For terrestrial plants, the majority of
excursion at the PTB, provides additional circumstantial evidence.
One possibility is that increased nutrient and particulate flux from
dominant incumbent genera and families at the PTB,
land into the marine system resulted in eutrophication and anoxia TJB and KPB did not become extinct, but rather were
leading to marine extinctions [71]. A possible parallel is observed at relegated to a lesser ecological role or were replaced by
the FFB (371 Mya) [84]. related species or genera. In marked contrast to the animal
It is noteworthy that a modern climate simulation study by
kingdom, these events have not had a primary role in the
Gedney et al. [85] hints at an alternative mechanism for changing
fluvial drainage patterns [85]. The study has attributed a 30% trajectory of plant evolution, having had little discernable
increase in global surface water run off, recorded since the 1960s to effect at higher taxonomic scales on plant origination,
a reduction in global evapotranspiration rates of terrestrial vegeta- extinction or in precipitating key morphological or ana-
tion, owing to increased stomatal closure in response to historical tomical innovations [6–9]. Plant clades demonstrate
increases in atmospheric CO2 concentration [85]. This result, if
incredible taxonomic resilience across faunal mass extinc-
correct, suggests that terrestrial vegetation can have a profound
impact on water run-off rates – and assumable fluvial drainage style tion boundaries perhaps owing to the great breadth of
– due to direct plant physiological responses to increased atmo- morphological, anatomical and physiological traits they
spheric CO2, in the absence of any significant change in ecological contain [6].
structure or composition of regional vegetation. This is a largely Despite this persistence, the fossil record of ancient
unexplored mechanism by which terrestrial plant physiological
responses to changes in atmospheric composition might have a
vegetation clearly demonstrates major structural and eco-
very significant impact on the marine system. Perhaps one of the logical upheaval before or coincident with peak faunal
most exciting avenues for future research, therefore, will involve extinctions [11–31] followed by prolonged periods of recov-
exploring further the linkage between terrestrial plant palaeoecolo- ery spanning thousands [73] to millions of years [10]. More
gical/physiological change and faunal biodiversity loss across importantly, these studies suggest that changes in ancient
extinction events of all magnitudes.
plant community ecology may have had a disproportionate
www.sciencedirect.com
 Review TRENDS in Ecology and Evolution Vol.22 No.10 555

influence on ecosystem function and faunal biodiversity 9 Traverse, A. (1988) Plant evolution dances to a different beat. Plant
and evolutionary mechanisms compared. Hist. Biol. 1, 277–356
[66–74]. These likely feedbacks have been largely over-
10 Looy, C.V. et al. (1999) The delayed resurgence of equatorial forests
looked because we have been using the wrong metric (first after the Permian-Triassic ecologic crisis. Proc. Natl Acad. Sci. U. S. A.
and last occurrences and taxonomic lists rather than 96, 13857–13862
abundance data) to measure ancient vegetation dynamics. 11 Looy, C.V. et al. (2001) Life in the end-Permian dead zone. Proc. Natl
In this respect we agree with conclusions from recent Acad. Sci. U. S. A 98, 7879–7883
12 Twitchett, R.J. et al. (2001) Rapid and synchronous collapse of marine
marine invertebrate studies [75] that indicate that biodi-
and terrestrial ecosystems during the end-Permian biotic crisis.
versity loss may not be the best metric for tracking ancient Geology 29, 351–354
ecosystem function [75]. 13 McElwain, J.C. et al. (2007) Macroecological responses of terrestrial
We are currently undergoing the sixth mass extinction vegetation to climatic and atmospheric change across the Triassic/
in Earth history and are faced with the task of ranking Jurassic boundary in East Greenland. Paleobiology 33 (4), 547–
573
species and ecosystem vulnerability to future global cli- 14 Wilf, P. and Johnson, K.R. (2004) Land plant extinction at the end of
mate change. Although the phylogenetic landscape was the Cretaceous: a quantitative analysis of the North Dakota
distinctly different in the geological past from that of the megafloral record. Paleobiology 30, 347–368
present day and therefore cannot inform species-based 15 Hotton, C. (2002) Palynology of the Cretaceous-Tertiary boundary in
predictions, a greater understanding of terrestrial veg- Central Montana: Evidence for extraterrestrial impact as a cause of the
terminal Cretaceous extinctions. In The Hell Creek Formation and the
etation dynamics during such periods of extreme environ- Cretaceous-Tertiary Boundary in Northern Great Plains: An Integrated
mental stress might make it possible to formulate broad Continental Record of the End Cretaceous (Hartman, J.H. et al., eds),pp.
ecological predictions for the future. For instance, 473–501, Geological Society of America Special Paper
increased extinction risk among reproductively specialized 16 Vajda, V. and McLoughlin, S. (2004) Fungal proliferation at the
Cretaceous-Tertiary boundary. Science 303, 1489
plant groups is a recurrent theme for the PTB, TJB and
17 Sweet, A.R. (2001) Plants, a yardstick for measuring the
KPB, providing empirical support, on evolutionary time- environmental consequences of the Cretaceous-Tertiary boundary
scales, for one prediction of modern metapopulation theory event. Geosci. Can. 28, 127–138
[65–77]. Model projections of future biodiversity response 18 Fowell, S.J. and Olsen, P.E. (1993) Time calibration of Triassic-Jurassic
to global climate change, based on modern plant distri- microfloral turnover, Eastern North-America. Tectonophysics 222,
361–369
butions alone, are limited by the difficulty of scaling from
19 Niklas, K.J. and Tiffney, B.H. (1994) The quantification of
individual species responses as independent and non- plant biodiversity through time. Philos. Trans. R. Soc. Lond. B
interacting entities to species responses within the eco- 345, 35–44
logical networks of communities [78]. In this respect, 20 DiMichele, W.A. and Phillips, T.L. (1996) Climate change, plant
palaeobiological research has a strong advantage, provid- extinctions, and vegetation recovery during the middle–late
Pennsylvanian transition: the case of tropical peat-forming
ing empirical evidence of how whole communities were environments in North America. In Biotic Recovery from Mass
shaped by natural global warming events of the past. Extinction (Hart, M.L., ed.), pp. 201–221, Geological Society of
Perhaps the past could be a key to our future? London
21 Gastaldo, R.A. et al. (1996) Out of the icehouse into the greenhouse: A
Late Paleozoic analog for modern global vegetational change. GSA
Acknowledgements
Today 6, 2–7
We thank I. Glasspool, R. Barclay and two anonymous reviewers for
22 Erwin, D.H. (1993) The Great Paleozoic Crisis, Columbia University
extremely helpful feedback and constructive criticism on the article.
Press
J. McElwain acknowledges funding from an EU Marie Curie Excellence
23 Xie, S.C. et al. (2005) Two episodes of microbial change coupled with
Grant, MEXT-CT-2006–042531 (MassExtinct). Funding for S. Punyasena
Permo/Triassic faunal mass extinction. Nature 434, 494–497
was provided by a US EPA STAR Fellowship FP-91637701–0.
24 Labandeira, C.C. and Eble, G.J. (in press) The fossil record of insect
diversity and disparity. In Gondwana Alive: Biodiversity and the
References Evolving Biosphere (Anderson, J. et al., eds), Witwatersrand
1 Gould, S.J. (1994) Tempo and mode in the macroevolutionary University Press
reconstruction of Darwinism. Proc. Natl Acad. Sci. U. S. A. 91, 25 Rees, P.M. (2002) Land-plant diversity and the end-Permian mass
6764–6771 extinction. Geology 30, 827–830
2 Sepkoski, J.J. (1982) Mass extinctions in the Phanerozoic oceans: a 26 McLoughlin, S. et al. (1997) Gondwanan floristic and sedimentological
review. Geol. Soc. Am. Spec. Pap. 191, 283–289 trends during the Permian-Triassic transition: new evidence from the
3 IPCC (2001) Climate Change 2001: The Scientific Basis. Contribution Amery Group, northern Prince Charles Mountains, East Antarctica.
of Working Group I to the Third Assessment Report of the Antarct. Sci. 9, 281–298
Intergovernmental Panel on Climate Change, Cambridge University 27 Retallack, G.J. (1997) Earliest Triassic origin of Isoetes and quillwort
Press evolutionary radiation. J. Paleontol. 71, 500–521
4 Benton, M.J. (1995) Diversification and extinction in the history of life. 28 Looy, C.V. et al. (2005) The ultrastructure and botanical affinity of
Science 268, 52–58 end-Permian spore tetrads. Int. J. Plant Sci. 166, 875–887
5 Niklas, K.J. (1985) Patterns in vascular land plant diversification: an 29 Visscher, H. et al. (2004) Environmental mutagenesis during the
analysis at the species level. In Phanerozoic Diversity Patterns: end-Permian ecological crisis. Proc. Natl Acad. Sci. U. S. A. 101,
Profiles in macroevolution (Valntine, J.W., ed.), pp. 97–128, 12952–12956
Princeton University Press 30 Foster, C.B. and Afonin, S.A. (2005) Abnormal pollen grains: an
6 Wing, S.L. (2004) Mass extinctions in plant evolution. In Extinctions outcome of deteriorating atmospheric conditions around the
in the History of Life (Taylor, P.D., ed.), pp. 61–97, Cambridge Permian–Triassic boundary. J. Geol. Soc. 162, 653–659
University Press 31 Visscher, H. et al. (1996) The terminal Paleozoic fungal event:
7 Kerp, H. et al. (2006) Typical Triassic Gondwanan floral elements in Evidence of terrestrial ecosystem destabilization and collapse. Proc.
the Upper Permian of the paleotropics. Geology 34, 265–268 Natl Acad. Sci. U. S. A. 93, 2155–2158
8 Knoll, A.H. (1984) Patterns of extinction in the fossil record of vascular 32 Eshet, Y. et al. (1995) Fungal event and Palynological record of
plants. In Extinctions (Nitecki, M.H., ed.), pp. 21–68, Chicago ecological crisis and recovery across the Permian-Triassic
University Press boundary. Geology 23, 967–970

www.sciencedirect.com
556 Review TRENDS in Ecology and Evolution Vol.22 No.10

33 Steiner, M.B. et al. (2003) Fungal abundance spike and the Permian- 56 McIver, E.E. (1999) Paleobotanical evidence for ecosystem disruption
Triassic boundary in the Karoo Supergroup (South Africa). at the Cretaceous-Tertiary boundary from Wood Mountain,
Palaeogeogr. Palaeoclimatol. Palaeoecol. 194, 405–414 Saskatchewan, Canada. Can. J. Earth Sci. 36, 775–789
34 Foster, C.B. et al. (2002) A revision of Reduviasporonites Wilson 1962: 57 Labandeira, C.C. et al. (2002) Impact of the terminal Cretaceous event
Description, illustration, comparison and biological affinities. on plant-insect associations. Proc. Natl Acad. Sci. U. S. A. 99, 2061–
Palynology 26, 35–58 2066
35 Breshears, D.D. et al. (2005) Regional vegetation die-off in response to 58 Wolfe, J.A. and Upchurch, G.R. (1986) Vegetation, climatic and floral
global-change-type drought. Proc. Natl Acad. Sci. U. S. A. 102, 15144– changes at the Cretaceous-Tertiary boundary. Nature 324, 148–152
15148 59 McIver, E.E. et al. (1991) 65000000-year-old fFlowers bearing pollen
36 Retallack, G.J. and Krull, E.S. (1999) Landscape ecological shift at the of the extinct Triprojectate complex - a Cretaceous Tertiary boundary
Permian-Triassic boundary in Antarctica. Aust. J. Earth Sci. 46, 785– survivor. Rev. Palaeobot. Palynol. 70, 77–88
812 60 Upchurch, G.R. and Wolfe, J.A. (1987) Extinction Patterns in
37 McElwain, J.C. et al. (1999) Fossil plants and global warming at the Laurales at the Cretaceous-Tertiary Boundary, Southern Rocky-
Triassic-Jurassic boundary. Science 285, 1386–1390 Mountains. Am. J. Bot. 74, 692–693
38 Retallack, G.J. (1995) Permian-Triassic extinctions on land. Science 61 McIver, E.E. and Bassinger, J.F. (1993) Fossil flora of the Plaeocene
267, 77–80 Ravenscrag Formation, southwestern Saskatchewan, Canada.
39 Retallack, G.J. et al. (1996) Global coal gap between Permian-Triassic Palaeontogr. Can. 10, 1–167
extinction and Middle Triassic recovery of peat-forming plants. Geol. 62 Tschudy, R.H. et al. (1984) Disruption of the terrestrial plant
Soc. Am. Bull. 108, 195–207 ecosystem at the Cretaceous-Tertiary boundary, Western Interior.
40 McLoughlin, S. and Drinnan, A.N. (1997) Fluvial sedimentology and Science 225, 1030–1032
revised stratigraphy of the Triassic Flagstone Bench Formation, 63 Wolfe, J.A. and Upchurch, G.R. (1987) Leaf assemblages across the
northern Prince Charles Mountains, East Antarctica. Geol. Mag. Cretaceous Tertiary boundary in the Raton Basin, New-Mexico and
134, 781–806 Colorado. Proc. Natl Acad. Sci. U. S. A. 84, 5096–5100
41 Hesselbo, S.P. et al. (2002) Terrestrial and marine extinction at the 64 Johnson, K.R. and Ellis, B. (2002) A tropical rainforest in Colorado 1.4
Triassic-Jurassic boundary synchronized with major carbon-cycle million years after the Cretaceous-Tertiary boundary. Science 296,
perturbation: A link to initiation of massive volcanism? Geology 30, 2379–2383
251–254 65 Gaston, K.J. (1994) Rarity, Chapman & Hall
42 Sepkoski, J.J. (1993) 10 Years in the Library - New Data Confirm 66 Roopnarine, P.D. (2006) Extinction cascades and catastrophe in
Paleontological Patterns. Paleobiology 19, 43–51 ancient food webs. Paleobiology 32, 1–19
43 Harris, T.M. (1937) The fossil flora of Scoresby Sound East Greenland. 67 Ward, P.D. et al. (2000) Altered river morphology in South Africa
Part 5: Stratigraphic relations of the plant beds. Meddelelser om related to the Permian-Triassic extinction. Science 289, 1740–1743
Gronland 112, 1–112 68 Michaelsen, P. (2002) Mass extinction of peat-forming plants and the
44 Hilton, J. and Bateman, R.M. (2006) Pteridosperms are the backbone effect on fluvial styles across the Permian-Triassic boundary,
of seed plant phylogeny. J. Torrey Bot. Soc. 133, 119–168 northern Bowen Basin, Australia. Palaeogeogr. Palaeoclimatol.
45 Olsen, P.E. et al. (2002) Ascent of dinosaurs linked to an iridium Palaeoecol. 179, 173–188
anomaly at the Triassic-Jurassic boundary. Science 296, 1305–1307 69 Sheldon, N.D. (2006) Abrupt chemical weathering increase across the
46 Pedersen, K.R. and Lund, J.J. (1980) Palynology of the plant-bearing Permian-Triassic boundary. Palaeogeogr. Palaeoclimatol. Palaeoecol.
Rhaetian to Hettangian Kap Stewart Formation, Scoresby Sund, East 231, 315–321
Greenland. Rev. Palaeobot. Palynol. 31, 1–69 70 Gastaldo, R.A. et al. (2005) Taphonomic trends of macrofloral
47 Visscher, H. and Brugman, W.A. (1981) Ranges of selected assemblages across the Permian-Triassic boundary, Karoo Basin,
Palynomorphs in the Alpine Triassic of Europe. Rev. Palaeobot. South Africa. Palaios 20, 479–497
Palynol. 34, 115–128 71 Sephton, M.A. et al. (2005) Catastrophic soil erosion during the end-
48 Vakhrameev, V.A. (1991) Jurassic and Cretaceous Floras and Permian biotic crisis. Geology 33, 941–944
Climates of the Earth, Cambridge University Press 72 Jablonski, D. (2005) Mass extinctions and macroevolution.
49 McElwain, J.C. et al. (2005) Changes in carbon dioxide during an Paleobiology 31, 192–210
oceanic anoxic event linked to intrusion into Gondwana coals. Nature 73 Arens, N.C. and Jahren, A.H. (2000) Carbon isotope excursion in
435, 479–482 atmospheric CO2 at the Cretaceous-Tertiary boundary: Evidence from
50 Fowell, S.J. and Traverse, A. (1995) Palynology and age of the Upper terrestrial sediments. Palaios 15, 314–322
Blomidon Formation, Fundy Basin, Nova-Scotia. Rev. Palaeobot. 74 Grice, K. et al. (2005) A potential biomarker for the Permian-Triassic
Palynol. 86, 211–233 ecological crisis. Earth Plan. Sci. Letts. 236, 315–321
51 Alvarez, W. (2003) Comparing the evidence relevant to impact and 75 Droser, M.L. et al. (2000) Decoupling of taxonomic and ecologic
flood basalt at times of major mass extinctions. Astrobiology 3, 153– severity of Phanerozoic marine mass extinctions. Geology 28, 675–
161 678
52 Nichols, D.J. and Johnson, K.R. (2002) Palynology and 76 Lawton, J.H. et al. (1994) Population-dynamic principles. Philos.
microstratigraphy of the K/T boundary in southwestern North Trans. R. Soc. Lond. B Biol. Sci. 344, 61–68
Dakota. In The Hell Creek Formation and the Cretaceous-Tertiary 77 McKinney, M.L. (1997) Extinction vulnerability and selectivity:
boundary in the Northern Great Plains: An integrated continental Combining ecological and paleontological views. Annu. Rev. Ecol.
record of the end of the Cretaceous (Hartman, J.H. et al., eds), pp. 95– Syst. 28, 495–516
144, Geological Society of America Special Paper 78 Walther, G.R. et al. (2002) Ecological responses to recent climate
53 Johnson, K.R. (2002) Megaflora of the Hell Creek and lower Fort change. Nature 416, 389–395
Union Formations in the western Dakotas: Vegetational response to 79 Wignall, P.B. (2001) Large igneous provinces and mass extinctions.
climate change, the Cretaceous-Tertiary boundary event, and rapid Earth Sci. Revs 53, 1–33
marine transgression. In The Hell Creek Formation and the 80 Burnham, R.J. et al. (1992) The reflection of deciduous forest
Cretaceous-Tertiary boundary in the northern Great Plains: An communities in leaf litter: implcations for autochthonous litter
integrated continental record of the end of the Cretaceous (Hartman, assemblages from the fossil record. Paleobiology 18, 30–49
J.H. et al., eds), pp. 329–391, Geological Society of America Special 81 Spicer, R.A. (1989) The formation and interpretation of plant fossil
Paper assemblages. Adv. Bot. Res. 16, 95–191
54 Johnson, K.R. et al. (2003) Overview of the Late Cretaceous, early 82 Steart, D.C. et al. (2005) Paleoecological implications of differential
Paleocene, and early Eocene megafloras of the Denver Basin, biomass and litter production in canopy trees in Australian
Colorado. Rocky Mount. Geol. 38, 101–120 Nothofagus and Eucalyptus forests. Palaios 20, 452–462
55 Vajda, V. et al. (2001) Indication of global deforestation at the 83 Smith, R.M.H. and Ward, P.D. (2001) Pattern of vertebrate
Creataceous-Tertiary boundary by New Zealand fern spike. Science extinctions across an event bed at the Permian-Triassic boundary
294, 1700–1702 in the Karoo Basin of South Africa. Geology 29, 1147–1150

www.sciencedirect.com
 Review TRENDS in Ecology and Evolution Vol.22 No.10 557

84 Algeo, T.J. and Scheckler, S.E. (1998) Terrestrial-marine 93 Kiehl, J.T. and Shields, C.A. (2005) Climate simulation of the latest
teleconnections in the Devonian: links between the evolution of land Permian: Implications for mass extinction. Geology 33, 757–760
plants, weathering processes, and marine anoxic events. Philos. Trans. 94 Kidder, D.L. and Worsley, T.R. (2004) Causes and consequences of
R. Soc. Lond. B 353, 113–128 extreme Permo-Triassic warming to globally equable climate and
85 Gedney, N. et al. (2006) Detection of a direct carbon dioxide effect in relation to the Permo-Triassic extinction and recovery. Palaeogeogr.
continental river runoff records. Nature 439, 835–838 Palaeoclimatol. Palaeoecol. 203, 207–237
86 Raup, D.M. and Sepkoski, J.J. (1982) Mass extinctions in the marine 95 Retallack, G.J. (2002) Carbon dioxide and climate over the past
fossil record. Science 215, 1501–1503 300 Myr. Philosophical Transactions of the Royal Society of London
87 Bambach, R.K. et al. (2002) Anatomical and ecological constraints on Series a-Mathematical Physical and Engineering Sciences 360,
Phanerozoic animal diversity in the marine realm. Proc. Natl Acad. 659–673
Sci. U. S. A. 99, 6854–6859 96 Hesselbo, S.P. et al. (2007) Triassic-Jurassic boundary events:
88 Palfy, J. et al. (2001) Carbon isotope anomaly and other geochemical Problems, progress, possibilities. Palaeogeography,
changes at the Triassic-Jurassic boundary from a marine section in Palaeoclimatology. Palaeoecology Special Issue 224, 1–423
Hungary. Geology 29, 1047–1050 97 Huynh, T.T. and Poulsen, C.J. (2005) Rising atmospheric CO2 as a
89 Willis, K.J. and McElwain, J.C. (2002) The Evolution of Plants, Oxford possible trigger for the end-Triassic mass extinction. Palaeogeogr.
University Press Palaeoclimatol. Palaeoecol. 217, 223–242
90 Raymond, A. and Metz, C. (1995) Laurussian land-plant diversity 98 Beerling, D.J. et al. (2001) Evidence for the recovery of terrestrial
during the Silurian and Devonian - mass extinction, sampling bias, or ecosystems ahead of marine primary production following a biotic
both. Paleobiology 21, 74–91 crisis at the Cretaceous-Tertiary boundary. J. Geol. Soc. 158, 737–
91 Krull, E.S. and Retallack, G.J. (2000) delta C-13 depth profiles from 740
paleosols across the Permian-Triassic boundary: Evidence for 99 Nordt, L. et al. (2002) Paleosol barometer indicates extreme
methane release. Geol. Soc. Am. Bull. 112, 1459–1472 fluctuations in atmospheric CO2 across the Cretaceous-Tertiary
92 Ward, P.D. and Garrison, G. (2003) Stable isotope results from the boundary. Geology 30, 703–706
Permian/Triassic boundary, South Africa, and Triassic/Jurassic 100 Wilf, P. et al. (2003) Correlated terrestrial and marine evidence for
boundary, Western Canada: Evidence against single strike impact global climate changes before mass extinction at the Cretaceous-
as causes for these two mass extinctions. Geochim. Cosmochim. Acta Paleogene boundary. Proc. Natl Acad. Sci. U. S. A. 100, 599–604
67, A525 ([email protected])

Five things you might not know about Elsevier

1.
Elsevier is a founder member of the WHO’s HINARI and AGORA initiatives, which enable the
world’s poorest countries to gain free access to scientific literature. More than 1000 journals,
including the Trends and Current Opinion collections and Drug Discovery Today, are now available
free of charge or at significantly reduced prices.

2.
The online archive of Elsevier’s premier Cell Press journal collection became freely available in
January 2005. Free access to the recent archive, including Cell, Neuron, Immunity and Current
Biology, is available on ScienceDirect and the Cell Press journal sites 12 months after articles are
first published.

3.
Have you contributed to an Elsevier journal, book or series? Did you know that all our authors are
entitled to a 30% discount on books and stand-alone CDs when ordered directly from us? For
more information, call our sales offices:

+1 800 782 4927 (USA) or +1 800 460 3110 (Canada, South and Central America)
or +44 (0)1865 474 010 (all other countries)

4.
Elsevier has a long tradition of liberal copyright policies and for many years has permitted both the
posting of preprints on public servers and the posting of final articles on internal servers. Now,
Elsevier has extended its author posting policy to allow authors to post the final text version of
their articles free of charge on their personal websites and institutional repositories or websites.

5.
The Elsevier Foundation is a knowledge-centered foundation that makes grants and contributions
throughout the world. A reflection of our culturally rich global organization, the Foundation has,
for example, funded the setting up of a video library to educate for children in Philadelphia,
provided storybooks to children in Cape Town, sponsored the creation of the Stanley L. Robbins
Visiting Professorship at Brigham and Women’s Hospital, and given funding to the 3rd
International Conference on Children’s Health and the Environment.

www.sciencedirect.com