The Lilliput Effect in The Aftermath of The End-Permian Extinction Event

Download as pdf or txt
Download as pdf or txt
You are on page 1of 13

Palaeogeography, Palaeoclimatology, Palaeoecology 252 (2007) 132 – 144

www.elsevier.com/locate/palaeo

The Lilliput effect in the aftermath of the


end-Permian extinction event
Richard J. Twitchett
School of Earth, Ocean and Environmental Science, University of Plymouth, Drake Circus, Plymouth, PL4 8AA, UK
Accepted 30 November 2006

Abstract

Early Triassic animal body fossils and trace fossils are small relative to those in older and younger intervals. Size decreases
sharply through the end-Permian extinction event and Permian/Triassic boundary, and the smallest sizes are encountered in the
parvus and isarcica Zones of the earliest Induan. Animals appearing within these two zones are also exceedingly small,
compared to younger congenerics and conspecifics. Temporary, dramatic size decrease of surviving taxa in the immediate aftermath
of the extinction event is an example of the Lilliput effect (coined by [Urbanek, A., 1993. Biotic crises in the history of Upper
Silurian graptoloids: a palaeobiological model. Historical Biology 7, 29–50.]). Body size increases somewhat from the carinata
Zone (mid-Induan) but remains depressed for the duration of the Early Triassic, and pre-extinction sizes are not commonly
recorded until at least the Middle Triassic. Marine and terrestrial faunas appear to be similarly affected. The Lilliput effect and
longer term size reduction could be the result of several factors. Environmental parameters such as marine anoxia, due to low
atmospheric concentrations of oxygen at this time coupled with sluggish ocean circulation in a greenhouse world, and food
shortage are the likely proximal causes for the Early Triassic Lilliput effect. No single cause can explain all the observations, and a
combination of factors are likely to be involved.
© 2007 Elsevier B.V. All rights reserved.

Keywords: Permian; Triassic; Body size; Fossil; Trace fossil

1. Introduction the Lilliput effect may therefore be crucial in understand-


ing the nature of ecological, environmental, and biolog-
A widespread, and yet little unstudied, evolutionary ical change during past biotic crises, especially during the
phenomenon is the “Lilliput effect”. This term, coined by immediate post-event aftermath.
Urbanek (1993), describes the pattern of size change Urbanek's (1993) initial study documented a size
through extinction events: in the immediate aftermath of decrease in several graptolite taxa in the wake of
such events fossil organisms are typically much smaller relatively small-scale biotic crises in the Late Silurian.
than during pre-extinction times. Body size is a key The Lilliput effect comprised one facet of his “post-
morphological variable, with implications for many event syndrome”, which affected the surviving taxa (the
aspects of an animal's biology, behaviour, and ecology “relic assemblage”). Fossil assemblages affected by this
(e.g. Barbault, 1988; Cotgreave, 1993). Understanding “post-event syndrome” are characterised by low diver-
sity, high abundance, and small body size (Urbanek,
E-mail address: rtwitchett@plymouth.ac.uk. 1993). Such assemblages are found in the immediate
0031-0182/$ - see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.palaeo.2006.11.038
R.J. Twitchett / Palaeogeography, Palaeoclimatology, Palaeoecology 252 (2007) 132–144 133

aftermath, or Survival Interval (sensu Kauffman and occurs after other biotic crises as well (e.g. in several
Erwin, 1995), of many extinction events, although molluscan taxa during the Pliensbachian–Toarcian event
Urbanek's (1993) terminology has rarely been specif- of the Early Jurassic; personal observation). It appears,
ically used in descriptions of such fossil assemblages. therefore, to be one factor that is common to all known
The Lilliput effect apparently occurred in the aftermath extinction crises and has important implications for our
of most of the major Phanerozoic extinction episodes, and understanding of the response of organisms to ecological
has been documented in a variety of animal groups, such as disturbance at both global and local scale. It may even
Early Silurian corals (Kaljo, 1996), Late Devonian allow us to predict the future response of the biosphere to
conodonts (Girard and Renaud, 1996; Renaud and Girard, present-day environmental change.
1999), and Early Danian echinoids (Jeffery, 2001), The mass extinction event that occurred in the latest
although the term “Lilliput effect” was not used by these Permian was the most severe event of the Phanerozoic
authors. There is growing evidence that the phenomenon (e.g. Erwin, 1993, 1994; Benton, 1995). It is ranked first
for both the magnitude of the diversity loss as well as for
the severity of the ecological impact on the marine and
terrestrial ecosystems of the Earth (McGhee et al., 2004).
It has long been recognised that the low diversity
assemblages of the Early Triassic aftermath are predom-
inantly composed of abundant, small-sized organisms
(e.g. Newell, 1952; Schubert and Bottjer, 1995), which are
typical attributes of Urbanek's (1993) “post-event syn-
drome”. Previous data have been a mixture of largely
qualitative or semi-quantitative studies. The aims of this
study are to assess the evidence for size decrease during
the Permian–Triassic extinction event, to document the
duration of the Lilliput effect, and to discuss some of the
potential causes of this phenomenon.

2. Small body size in the Early Triassic

Newell (1952) was one of the first to note in print that


Early Triassic fossils are small, although many subse-
quent authors have come to the same conclusion since
(e.g. Hayami, 1997, 1998). Certainly, the surprisingly
small size of the fossils is one of the most striking
aspects of collecting in Lower Triassic strata, and this
observation does not change over successive field
seasons. Despite more than a decade of collecting and
museum studies worldwide, I have yet to find an Early
Triassic bivalve more than 80 mm in size, or an Early
Triassic gastropod more than 40 mm, and most are
substantially smaller.
Such qualitative observations are also supported by
more semi-quantitative data. Hayami's (1998) study of
bivalve size through time, based on analysis of data in the
Fig. 1. Preliminary data of size change through the Permian and Triassic Treatise on Invertebrate Paleontology, showed that the
in gastropods (A) and bivalves (B). Square symbols show mean average geometric mean size of Early Triassic bivalve
geometric size; vertical bars indicate one standard deviation above and genera (25.8 mm) is smaller than genera from the Middle
below the mean; dashed line shows maximum geometric size. Data were Triassic (27.2 mm) and Early Permian (27.0 mm),
collected by measuring fossil specimens in the collections of the Natural
although because of the way the data are presented it is
History Museum London, supplemented by measurements made on
specimens in the field. Data are recorded in Table 1. Full data set is unclear whether these differences are statistically signif-
available from the author on request. C = Cisuralian; G = Guadalupian; icant. The data presented in Hayami (1998) do not,
L = Lopingian; E = Early; M = Middle; L = Late. however, show a reduction in mean size across the
134 R.J. Twitchett / Palaeogeography, Palaeoclimatology, Palaeoecology 252 (2007) 132–144

Table 1
Geometric sizes of bivalve and gastropod specimens through the Permian–Triassic interval
Permian Triassic
Cissuralian Guadalupian Lopingian Early Middle Late
Gastropods
Mean (mm) – 33.0 17.9 4.9 10.5 29.9
s.d. (mm) – 13.0 6.9 2.6 7.4 20.2
Maximum (mm) – 49.1 48.7 17.0 39.0 83.5
n – 30 54 153 83 47

Bivalves
Mean (mm) 31.2 – 21.3 12.1 22.3 23.4
s.d. (mm) 11.2 – 7.3 0.9 13.0 17.3
Maximum (mm) 58.8 – 42.4 80 47 75.2
n 14 – 108 1166 27 137

Permian/Triassic (P/Tr) boundary, as was claimed in a Similar patterns are seen in all other benthic groups
previous publication (Hayami, 1997), and Late Permian studied thus far. For example, Early Triassic ophiuroids are
genera are smaller (mean size 23.7 mm) than Early much smaller (maximum disk diameter of 10 mm) than
Triassic taxa. However, the proportion of very large-sized Middle Triassic (20 mm), modern temperate (30 mm), or
genera (i.e. those in excess of 64 mm) does appear to modern tropical (45 mm) taxa (Twitchett et al., 2005).
decrease across the P/Tr boundary (Hayami, 1998, p. 39). Small body size is also characteristic of Early Triassic
Other molluscan taxa are also smaller in the Early marine and terrestrial vertebrates (e.g. Smith, 1995;
Triassic. “Microgastropods”, which are defined as being Tverdokhlebov et al., 2002), although quantitative data of
gastropods that are less than 10 mm in height (Fraiser and size change are currently lacking. Early Triassic terrestrial
Bottjer, 2004), are so common in Lower Triassic rocks ecosystems are typically described as being dominated by
that they may reach rock-building densities (Fraiser et al., small organisms, at least in the very earliest Triassic. In
2005). Early Triassic gastropod assemblages in all known Russia, small temnospondyls characterise the beginning of
marine environments are dominated by microgastropod the Triassic, and the ecological niches for large herbivores
species, whereas Permian, Middle Triassic, or modern and predators were unoccupied (Tverdokhlebov et al.,
gastropod assemblages tend to be dominated by larger 2002). Similarly, in South Africa it appears that the
gastropods (Fraiser and Bottjer, 2004). In a parallel study terrestrial ecosystems of the Early Triassic were dominated
based on literature review, Payne (2005) also demonstrat- by small sized tetrapods (Smith, 1995).
ed a similar pattern of size change in Permian–Triassic These qualitative and quantitative data give a first order
(P–Tr) gastropods: namely, size decrease into the Early understanding of P–Tr size change at the stage or series
Triassic followed by size increase in the Middle Triassic. level. All animal groups apparently record identical
As a complementary study to the literature-based changes, with size minima in the Early Triassic. However,
compilations by Payne (2005) and Hayami (1998), I am a few studies have also documented size changes at a much
compiling data from measurements based on actual higher resolution (i.e. sub-stage or zonal scale). The
fossil specimens from field studies and museum earliest such study involved analyses of the diameter of
collections. Preliminary results are shown here plotted trace fossil burrows from the Bellerophon and Werfen
at series scale (Fig. 1, Table 1) and the patterns of size formations of northern Italy (Twitchett, 1999). From the
change through the Permian and Triassic are very Late Permian Bellerophon Formation into the parvus and
similar to those found in previous studies. Both the isarcica conodont zones (i.e. early–mid Griesbachian,
mean and maximum body sizes of gastropods decrease earliest Induan) of the Werfen Formation, burrow diameter
through the Permian to reach a minimum in the Early decreased by an order of magnitude. As burrow diameter is
Triassic, before increasing during the subsequent a good proxy of the size of the organism that made the
Triassic (Fig. 1A). The mean body size of bivalves burrow (e.g. Savrda et al., 1984; Savrda and Bottjer, 1986),
also reaches a minimum in the Early Triassic, although these data imply a dramatic reduction in the body sizes of
the maximum size (e.g. specimens of Claraia from the the infaunal, soft-bodied, trace-making community.
Perth Basin of western Australia) exceeds that of the Since this initial study, further field measurements have
Lopingian. been made through the same succession – the data set now
R.J. Twitchett / Palaeogeography, Palaeoclimatology, Palaeoecology 252 (2007) 132–144 135

Fig. 2. Burrow diameters of trace fossils through the Late Permian to Middle Triassic of the Dolomites, northern Italy. A: all burrow types; B: vertical
domichnia of suspension feeders (i.e. Arenicolites, Skolithos, Diplocraterion); C: fodinichnia of deposit feeders (i.e. Planolites). Circles indicate
mean value; vertical lines indicate one standard deviation above, and below, the mean; shaded region indicates total burrow range between the
maximum and minimum recorded sizes. Measurements recorded in the field to the nearest 0.5 mm. TOH = Tesero Oolite Horizon; AzH = Andraz
Horizon; GOM = Gastropod Oolite Member.

comprises some 1257 samples – yet the patterns remain San Lucano Member, but it is not until the Anisian that
robust (Fig. 2). The initial decrease in both maximum and burrow size commonly exceeds 10–15 mm, with
mean burrow size from the Bellerophon Formation into the Thalassinoides reaching 30 mm and Rhizocorallium
Mazzin Member of the Werfen Formation remains the reaching 20 mm in diameter.
most severe decline. Following this there is a gradual Recent work on the shelly body fossils of the Late
increase in mean burrow size through the Early Triassic, Permian and Early Triassic record of northern Italy has
interrupted by temporary decreases in the upper Siusi revealed similar patterns of size change (Price-Lloyd
Member and Campil Member. Similar patterns are and Twitchett, 2002) (Fig. 3). The taxa Bellerophon and
observed when the domichnia of suspension feeders (e.g. Lingula show statistically significant reductions in body
Arenicolites, Skolithos) and the fodinichnia (e.g. Planolites) size from the Bellerophon Formation through the
are considered separately (Fig. 2). One or two rare, large extinction event into the Mazzin Member of the Werfen
burrows are encountered in the lower Siusi Member and Formation. In addition, taxa which appear in the
136 R.J. Twitchett / Palaeogeography, Palaeoclimatology, Palaeoecology 252 (2007) 132–144

immediate aftermath of the event, namely Claraia and Event. The taxon P. dubius spans the entire extinction–
Unionites, are also smallest in the Mazzin Member. recovery interval, although the Lilliput forms are assigned
All taxa show significant increases in body size in the to the subspecies P. dubius parvus, whereas the ‘normal’
lower Siusi Member. pre-extinction, and later recovery, forms are assigned to the
subspecies P. dubius frequens.
3. The Early Triassic Lilliput effect Adhering to Urbanek's original definition of the
Lilliput effect, it is clear that the only Permian–Triassic
Urbanek's (1993) definition of the Lilliput effect data described above that could fit with this definition are
describes it as “the occurrence of diminuative forms the statistically significant, short-term size changes within
among some of the species in the relic assemblages” of the the surviving taxa Lingula and Bellerophon during the
immediate post-extinction aftermath. In other words, the Griesbachian (early Induan), which were first described
Lilliput effect, as originally defined, describes size by Price-Lloyd and Twitchett (2002). Poor preservation
reduction within specific taxa (species) that survive the and dramatic size change mean that the Lilliput forms of
extinction event and is a temporary phenomenon confined these taxa in the Dolomites sections are typically not
to the Survival Interval (sensu Kauffman and Erwin, 1995). assigned to species, but are usually referred to as Lingula
The classic example of the Lilliput effect described by sp. or Bellerophon sp. Until further taxonomic study is
Urbanek (1993) is the mass occurrence of short and slim undertaken, it is not possible to be certain that size
Pristiograptus dubius in the aftermath of the lundgreni reduction is happening at the specific or generic level.
Regardless, it is clear that the remaining data
concerning size reduction in the Early Triassic (documen-
ted above) do not comfortably fit within Urbanek's (1993)
definition of the Lilliput effect: small size is recorded in
newly originating taxa as well as surviving taxa; in the
relative proportions of ‘small’ and ‘large’ species within
higher taxa; and through the whole of the Early Triassic,
not just the immediate post-extinction aftermath. This
represents a larger-scale, longer-term depression of body
size, affecting the entire fauna (survivors as well as newly
originating taxa) and lasting through the Survival Interval
and well into the Recovery Interval. This is not the Lilliput
effect in the sense of Urbanek (1993) and may or may not
be related to the extinction event. The apparent long-term
decline in body size through the Permian (e.g. Fig. 1) may
be related to long term environmental changes (see below)
or successive extinction events (end-Guadalupian, end-
Permian). Typically, studies based largely on literature
review and/or semi-quantitative analyses of taxa at the
generic or family level (e.g. Payne, 2005) will highlight
this longer term depression of body size, whereas
measurements of individual, well-dated specimens from
field or museum collections are necessary to document the
Fig. 3. Size change in benthic marine invertebrates in the aftermath of shorter term Lilliput effect. Regarding trace fossils: size
the end-Permian extinction event. Black square = Bellerophon (n = 42); reduction within specific ichnogenera or ichnospecies in
grey circle = Lingula (n = 195); black diamond = Claraia (n = 721); grey the immediate aftermath of an extinction event is herein
triangle = Unionites (n = 258). Symbols show mean size; horizontal
lines indicate maximum and minimum recorded sizes. Size calculated considered to also represent the Lilliput effect, although it
as the geometric mean of length and height. All within-taxon size is recognised that a single ichnotaxon may have been
changes are significant at the 5% level (Kolmogorov–Smirnov test). produced by more than one animal species. Studies
Data collected in the field from localities in northern Italy showing a size reduction of the entire ichnofauna, without
(Bellerophon and Werfen formations). The Griesbachian is divided identifying the individual ichnotaxa (e.g. Twitchett, 1999;
into three conodont zones: parvus, isarcica, and carinata. Arrow
indicates approximate position of extinction event. Data from Price-
Fig. 1), do not demonstrate the Lilliput effect in the strict
Lloyd and Twitchett (2002). Bell Fm = Bellerophon Formation; sense, but do record this longer term depression of body
To = Tesero Oolite Horizon; Mb = Member; Az = Andraz Horizon. size.
R.J. Twitchett / Palaeogeography, Palaeoclimatology, Palaeoecology 252 (2007) 132–144 137

4. Causes of size reduction and the Lilliput effect

4.1. Evolutionary patterns

The Lilliput effect and the longer-term Early Triassic


size reduction are recorded at different taxonomic
scales, are of different durations and may be the result
of similar, or entirely different, processes. Questions
concerning the causes of post-extinction size decrease
can also be addressed in several ways. The nature of
these hypotheses are quite different, with the larger scale
incorporating preservational and taphonomic effects as
well as (potentially) macroevolutionary factors, whereas
at the scale of an individual species, the biological
responses of individual organisms to local environmen- Fig. 4. Simplified model showing the effect of a selective extinction
tal change are probably most important. Urbanek (1993) event on the size distribution of the taxa in the immediate extinction
aftermath. Horizontal lines represent the hypothetical fossil ranges of
discussed the ecophenotypic response of individuals to
taxa at a number of different hypothetical sizes. Note a decrease in
palaeoenvironmental changes in parameters such as maximum and mean size following the event.
temperature, salinity and food supply, but concluded
that competing alternatives could not be tested from
fossil remains. However, regarding the Permian– susceptible to extinction for a variety of reasons including
Triassic interval, it may be possible to reject at least their greater energy requirements, longer generation
some of these competing possibilities (see below). times, and relatively lower population sizes (Labarbera,
Finally, one significant factor in the interpretation of the 1986; Barbault, 1988; Cotgreave, 1993; Jablonski, 1996).
observed changes in size is the nature of the data, Hayami (1997) invoked size selection as an explanation
including both the taxonomic level of analysis and the for size changes in Cretaceous–Tertiary (K–T) and P–Tr
parameter under consideration: is a reduction being bivalves. Regarding the K–T event, the communities of
recorded in maximum size, minimum size, and mean small vertebrates or planktonic foraminifera of the earliest
size, or in one or two, but not all three, of these Danian have also been previously interpreted as being the
parameters? result of selection against large-bodied forms at this time
(see Jablonski, 1996 and refs. therein).
4.1.1. Extinction of large taxa? Recent data have, however, cast doubt on the size-
A seemingly simple explanation for post-extinction selection hypothesis. At the Cretaceous/Tertiary (K/T)
size reduction is the extinction of large-bodied (K-selected) boundary event, while all of the large, specialist (K-
taxa leaving only small-bodied survivors. Assuming that selected) planktonic foraminifera disappeared, so did all
some time would be required for new large-sized taxa to the small trochospiral ecological generalist taxa (except the
evolve from small-sized survivors, extinction of large taxa hedbergellids, which became extinct in the early Danian)
would result in a reduction in both the maximum and mean (Keller, 2003). While very large taxa are demonstrably
size of taxa present in the immediate extinction aftermath, absent from Danian assemblages, the taxa that became
but not necessarily any change in the minimum size extinct were not exclusively large. Regarding the
(Fig. 4). However, simple extinction of large taxa is not a vertebrates, Jablonski (1996) has argued that the K–T
potential explanation of the Lilliput effect, as short-term, data are the result of clade-specific rather than size-specific
post-extinction, within-lineage size decrease in surviving selectivity: all large Maastrichtian land animals were “non-
(i.e. small) species would not be predicted to occur. Loss of avian dinosaurs”, and all “non-avian dinosaurs” (both large
large taxa is recorded at many events and in many groups and small) became extinct. Data on marine invertebrates
of organisms, such as during the end-Cretaceous event (bivalves and gastropods) show no size difference between
where it is responsible for a difference in the maximum survivors and victims (Jablonski and Raup, 1995;
sizes of pre- and post-event planktonic foraminifera. Jablonski, 1996). A recent study of bivalves through the
A more intriguing problem is whether extinction of K–T, mid-Eocene, and Eocene–Oligocene events also
large taxa is the result of random chance or the result of found no evidence that the extinctions were size-selective
active size selection against large taxa during major at the genus level (Lockwood, 2005). Likewise, a
extinction events. Large-bodied animals should be more comprehensive study of K–T echinoids also shows that
138 R.J. Twitchett / Palaeogeography, Palaeoclimatology, Palaeoecology 252 (2007) 132–144

reduction in the mean, and possibly minimum, body size


of taxa in the post-extinction assemblages, but the
maximum body size should remain more or less the
same. Likewise, within-species size reduction in
surviving taxa is not predicted, and so this mechanism
cannot explain the Lilliput effect. Two factors could be
involved: (1) the origination of new taxa and/or (2) the
appearance of pioneering opportunists.
Stanley (1973) observed that, generally, animal taxa
tend to originate at small size. Extinction events involve
widespread loss of biodiversity, presumably leaving many
vacant niches, which should promote the appearance and
radiation of many new groups. Therefore, if Stanley's
(1973) hypothesis is correct, the aftermath of an extinction
Fig. 5. Simplified model showing the effect on post-extinction origination event should be an interval with above average numbers
of many small taxa on the size distribution of the post-extinction
community. Horizontal lines represent hypothetical fossil ranges of taxa at
of newly originating, small species. The founders of these
a number of different hypothetical sizes. Note that the extinction is random new groups have been termed “crisis progenitors” and do,
with respect to size and that the maximum size is unaffected. indeed, tend to be small sized (Kauffman and Harries,
1996). To date, the only exception to this general rule is in
the aftermath of the Eocene–Oligocene event, where
size selection played no part in the extinction event (Smith newly originating bivalve genera were apparently larger
and Jeffrey, 1998), although a number of surviving than surviving taxa (Lockwood, 2005).
lineages did become smaller (the Lilliput effect) (Jeffery, An alternative is that extinction events cause an
2001). While some examples of post-extinction size explosion in pioneering opportunists, i.e. long ranging
increase have been explained in terms of changes in life- taxa that were already present locally prior to the event but
history strategy (e.g. Hallam, 1975), to date no study has which undergo rapid expansion in numbers (blooms) in
yet demonstrated that extinction events themselves are the disturbed environments of the immediate aftermath
size-selective. (Kauffman and Harries, 1996). Communities dominated
In summary, while it is obvious that large-bodied taxa by opportunists occur in the wake of modern, smaller-
do become extinct, and are therefore missing from the scale, defaunation events, following removal of the
post-extinction communities, no quantitative study has previous climax community. Some authors also regard
yet demonstrated that larger taxa are preferentially lost. It opportunism as an important survival strategy during time
should be noted, however, that most quantitative studies of mass extinction (Harries et al., 1996), and the
documented to date have involved the K–T event and few appearance of post-extinction opportunists may therefore
quantitative tests of size selection have hitherto been be expected. Opportunists are often described as being r-
published for other extinction intervals. Extinction of selected taxa (i.e. animals whose reproductive strategy
large-bodied animals during the Late Permian event is involves early maturation and the production of a large
considered to be a contributory factor in the long term number of offspring). A common trait of such taxa is
post-extinction depression of body size during the Early small size, and if all Early Triassic environments were
Triassic, but it is not a cause of the short-term Lilliput disturbed, and if all communities were dominated by
effect. Likewise, loss of large taxa during the end- opportunists, then this could be an explanation for the
Guadalupian event may have contributed to the decline in smaller than expected body sizes encountered. Certainly,
maximum and mean size of, for example, gastropods Fraiser and Bottjer (2004) interpreted the microgastro-
through the latter half of the Permian (Fig. 1A). pods of the Early Triassic as opportunists. However, it is
often difficult to identify the age of maturation and/or the
4.1.2. The post-crisis appearance of many small taxa? size and number of offspring produced in fossil examples
Regardless of whether the extinction episode was and small size alone is not an unequivocal indicator of
selective against large animals, the fact that Early opportunism. For example, trace fossil evidence from
Triassic fossils appear to be smaller than expected may Early Triassic marginal marine facies of the western USA
be due to the appearance of a large number of small taxa and northern Italy indicates the presence of persistent
in the extinction aftermath (Fig. 5). Unlike the previous climax communities of small-sized ophiuroids (Twitchett
hypothesis (Fig. 4), this pattern of size change involves a et al., 2005).
R.J. Twitchett / Palaeogeography, Palaeoclimatology, Palaeoecology 252 (2007) 132–144 139

extinction episodes and in a number of groups such as P–


Tr gastropods (Erwin, 1996), sponges, and bivalves
(Twitchett, 2001). Can the Lilliput effect be explained by
a size bias in the Lazarus effect?
Such a bias has yet to be demonstrated. Temporary
disappearance of specific taxa from the fossil record may
be caused in a number of different ways, including basic
rock-record bias such as the temporary absence of
fossiliferous rocks of the correct facies (Paul and
Donovan, 1998; Smith, 2001). Certain marginal environ-
ments, such as brackish estuaries or hypersaline lagoons,
tend to be dominated by small animals. If the Lower
Triassic rock record comprised many more examples of
facies deposited in such settings than the Lopingian or
Fig. 6. Simplified model showing the effect of a temporary later Triassic records, then the sampled fossil record could
disappearance of large sized taxa on the size distribution of the post-
extinction community. Horizontal lines represent hypothetical fossil
be biased towards smaller-sized organisms. Burrow size
ranges; dashed lines indicate Lazarus taxa. Note that the extinction is minima in the upper Siusi Member and Campil Member
random with respect to size and the preserved taxa (solid lines) in the of the Werfen Formation (Fig. 2) are attributed to the
extinction aftermath show a reduction in maximum and mean body marginal, brackish depositional environment of these
size with respect to the pre-event community. units (Twitchett, 1999). Other possible biases include
changes in the nature of fossil preservation: for example,
Data from the Lower Triassic Werfen Formation of Erwin's (1996) hypothesis that a dearth of Early Triassic
northern Italy show that the bivalve taxa Unionites and silicified faunas may be to blame for the P–Tr Lazarus
Claraia, which appear in the immediate aftermath of the effect in gastropods, as early silicification can help
biotic crisis, are indeed smaller than later examples of the preserve aragonitic shells, which otherwise may dissolve
same genera (Fig. 3). Microgastropods, which may or may prior to fossilisation.
not be opportunists, are also very widespread throughout Another possibility is that population size determines
most of the Early Triassic (Fraiser et al., 2005). However, it whether or not a particular species will be represented in
is also clear from studies conducted to date that there was the fossil record (Marshall, 1998; Wignall and Benton,
undoubtedly a reduction in the maximum size of Early 1999). This model assumes that the chance of any
Triassic taxa (e.g. Payne, 2005) (cf. Fig. 5). Thus, while the particular individual entering the fossil record is
widespread appearance of small taxa (either newly vanishingly small, so only the most populous taxa will
originating species or blooms of opportunists) in the be fossilised. Populations of living organisms fluctuate
extinction aftermath may be a contributory factor, it is not in response to ecological and environmental events.
the sole explanation for the observed size changes in P–Tr Should numbers temporarily fall below the threshold for
animal groups. entering the fossil record, then that particular taxon will
become a Lazarus taxon until population size recovers
4.1.3. The temporary disappearance of large taxa? once more (Twitchett, 2001; see also discussion in Fara,
It is possible that the absence of large taxa through 2001). As large-bodied animals generally have lower
most of the Early Triassic is not a real biological population sizes than smaller ones, they may be more
phenomenon at all, but may be an artefact of the fossil susceptible to falling below this fossilisation threshold.
record, and caused by differences in the preservation It is assumed that an extinction event, if it were to have
potential of large and small taxa through an extinction any affect on the numbers of living animals, would
interval. If large taxa temporarily disappear from the fossil reduce, rather than increase, population sizes.
record, then only small taxa will be found as fossils, and Given the complexity of the processes leading to
thus both the mean and maximum body size of the post- fossilisation, in particular time averaging, assessing the
extinction fossil assemblage will be reduced (Fig. 6). absolute (or even relative) population sizes of extinct
Taxa that temporarily disappear from the fossil record organisms from their fossil remains is not possible.
are termed “Lazarus taxa”. This term was coined to Therefore, the assumptions of the population-dependent
describe the temporary disappearance of taxa through model are very difficult to test (see discussion in
extinction events (Flessa and Jablonski, 1983) and the Twitchett, 2001). However, the possibility that there is a
Lazarus effect has been documented through all major bias in the Lazarus effect between large and small taxa
140 R.J. Twitchett / Palaeogeography, Palaeoclimatology, Palaeoecology 252 (2007) 132–144

millions of years. Suboptimal environmental conditions


may lead to slow growth rates and stunting, or may
promote heterochronic changes such as an earlier onset of
maturity (progenesis). Analysis of growth lines within
shelly taxa could be used to explore these alternatives
further, but no such study has yet been published for any
taxon through any extinction interval. Suboptimal envi-
ronmental conditions might also place upper limits on the
size of newly originating species. High juvenile mortality
in the wake of the extinction event is rejected as a possible
explanation as it may reduce the mean and minimum size
of fossilised individuals but not the maximum size. Indeed,
in all discussions concerning the Lilliput effect and size
reduction the assumption is made that we are dealing with
Fig. 7. Simplified model showing the effect of a temporary reduction in
changes in the adult forms. Likewise, although it is
the body sizes of surviving taxa on the size distribution of the post-
extinction community. Horizontal lines represent hypothetical fossil possible that different taxonomic groups were responding
ranges of taxa at a number of different sizes. Note that the extinction is to different environmental pressures, the fact that size
random with respect to size and there is an overall reduction in reduction affects both shelly taxa and non-mineralised
maximum, mean and minimum size. burrowing infauna (Figs. 1–3) strongly implies that it was
not solely caused by a problem with skeletal secretion or
(Fig. 6) can be tested, and if it exists would be a possible biocalcification. Finally, when dealing with fossilised
explanation for the apparent absence of large individuals remains it is best to bear in mind other potential causes
and large taxa in the aftermath of the end-Permian mass such as biases in fossilisation. For example, there may be
extinction event. Erwin (1996) demonstrated that the an under-representation of large individuals through rarity
Lazarus effect in gastropods increases from the Middle (c.f. hypothesis 3 above). As discussed by Urbanek (1993),
Permian through to the Early Triassic, which matches the genetic changes are probably not responsible and so are not
pattern of gastropod size change recorded during this considered further here.
interval (Fig. 1). Possible evidence of within-lineage size decrease
(temporary stunting) associated with the end-Permian
4.1.4. Within-lineage size decrease? extinction event is given by studies of Bellerophon and
A fourth alternative is that surviving taxa all underwent Lingula from northern Italy (Fig. 3). However, unre-
a reduction in body size in the aftermath of the extinction solved taxonomic issues mean that it is not possible to
event such that fossilised individuals of particular species be certain that size reduction is happening at the species
are smaller in the post-extinction interval than during the or generic level. The strict definition of the Lilliput
pre-extinction times (Fig. 7). Maximum, mean and effect (sensu Urbanek, 1993) implies that size reduction
minimum sizes would all be reduced. The Lilliput effect, is occurring within species (however defined).
as defined by Urbanek (1993), represents temporary size
decrease within surviving species that lasted on the order 4.1.5. All of the above?
of a few hundred thousand years. If such a size decrease The four alternative explanations for the Lilliput effect,
lasted for a much longer interval of time (i.e. many outlined above, can all be distinguished with fossil data, as
millions of years) then it could also be a contributory the predicted patterns of size distribution are different in
factor to the lengthy post-extinction body size low that all cases (Figs. 4–7). Of course, it is quite possible that
lasts through the entire Early Triassic. different patterns will be observed in different taxonomic
There are several possible ways to achieve size groups during the same extinction-recovery event, that
reduction within a fossil taxon. Ecophenotypic responses different events may record different patterns of size
to changes in parameters such as temperature, salinity, distribution, or that some taxa may show a combination of
oxygen levels, and food supply are most likely, and the patterns (c.f. Lockwood, 2005). During the P–Tr interval,
latter two are more fully explored below (see also Hallam, small size characterises both survivors and newly
1965; Lockwood, 2005; Twitchett, 2006). The Lilliput appearing taxa and thus involves at least a combination
effect would be caused by relatively short term changes, of hypotheses (2) and (4) above. A thorough, quantitative
whereas longer term reductions in size would imply test of hypotheses (1) and (3) has yet to be attempted for
suboptimal environmental conditions lasting many this extinction event.
R.J. Twitchett / Palaeogeography, Palaeoclimatology, Palaeoecology 252 (2007) 132–144 141

4.2. Environmental controls on body size this is most often demonstrated by measuring the size of
burrows that these animals leave behind (e.g. Savrda and
Stanley (1973) hypothesised that for any given set of Bottjer, 1986). Finally, experimental data generally show
environmental parameters there is an optimal body size that animals reach smaller maximum adult size, and/or
for animals, and that size trends result from the process of mature at smaller size, under conditions of sub-lethal
attaining optimal size in the face of changing environ- hypoxia (e.g. in copepods; Richmond et al., 2006).
mental conditions. If the aftermaths of mass extinction Regarding the data from northern Italy (Figs. 2 and 3),
events represent times when environmental conditions are the smallest burrow diameters and smallest shelly
such that small size is the optimum, then this would invertebrates are found in the Mazzin Member of the
explain both within-lineage size decrease and the small Werfen Formation. This unit was deposited under very
sizes of newly appearing taxa. While many environmental low, but fluctuating, oxygen conditions as indicated by
factors may result in size decrease (Hallam, 1965), two are facies analysis and geochemistry (Wignall and Twitchett,
considered most important with regards to explaining the 1996, 2002). In localities of identical age, but where facies
Lilliput effect in the aftermath of the end-Permian indicate well oxygenated conditions, such as in Oman
extinction event: anoxia and food shortage. (Krystyn et al., 2003), diversity is much higher and the
sizes of shelly taxa (e.g. gastropods) are also slightly
4.2.1. Anoxia larger (Twitchett et al., 2004; Wheeley and Twitchett,
A substantial body of evidence has accumulated that 2005). Thus, from these preliminary data, the correspon-
demonstrates that anoxic, and even euxinic, conditions dence between body size decrease and anoxia is good,
were widespread in most marine environments and suggesting that it may be a possible cause of the Lilliput
regions through most of the Early Triassic (e.g. Wignall effect in the parvus and isarcica zones of the Induan. The
and Twitchett, 1996, 2002; Grice et al., 2005). In the deep very low atmospheric oxygen levels that apparently
ocean, oxygen restriction lasted from the Lopingian to characterised the entire Early Triassic (Berner, 2005) may
Middle Triassic (e.g. Isozaki, 1997), whereas in some be a cause of the longer-term depression of body sizes
shallow shelves marine anoxia is only recorded in the observed in all animal groups studied to date.
early Induan (e.g. Wignall and Twitchett, 1996). Taking a However, while there is some correspondence between
global view, peak anoxia (i.e. the widest geographic low oxygen levels and small size in these data, it also
spread of oxygen-poor waters) occurred in the parvus and seems apparent that anoxia alone cannot account for all
isarcica zones of the early Induan (Wignall and Twitchett, the data thus far collected. In northern Italy, the Lilliput
2002). At this time, only shallow seamounts and offshore effect lasts until the latest Griesbachian and sizes begin to
highs were free of evidence of marine anoxia (e.g. increase in the lower Siusi Member of the Werfen
Krystyn et al., 2003). Low oxygen conditions were Formation (Figs. 2 and 3). The lower Siusi Member,
probably the direct result of global warming and a however, is still affected by low oxygen conditions to
subsequent sluggishness in oceanic circulation, and have some extent (e.g. Wignall and Twitchett, 1996). Also,
been recreated in computer simulation models (Hotinski there are significant regional differences in the sizes of
et al., 2001). High temperatures would have also have some taxa that cannot be explained by differences in
caused a decrease in the amount of dissolved oxygen in oxygen levels. Measurements of the sizes of individual
the oceans, exacerbated by the relatively low concentra- Claraia specimens from the basal Wordie Creek Forma-
tion of oxygen within the atmosphere (Berner, 2005). tion of East Greenland and the Hovea Member of the
A large number of published data, from both modern Kockatea Shale Formation of Western Australia show that
experimental studies and analyses of the fossil record, the Australian specimens are significantly larger (Fig. 8).
have demonstrated the link between oxygen concentra- Yet, all of these specimens, from both regions, derive from
tion and body size. Episodes of surprisingly large body typical Griesbachian oxygen-restricted facies, charac-
size in terrestrial animals correlate with times of high terised by dark grey, parallel-laminated mudstones with
atmospheric oxygen and the inferred cause–effect link has rare, bioturbated horizons that were formed under similar,
been strengthened by modern experiments, such as in fluctuating low oxygen conditions (e.g. Wignall and
regard to Carboniferous dragonflies (Graham et al., 1995; Twitchett, 2002; Grice et al., 2005). Thus, the size
Harrison and Lighton, 1998). Likewise, in marine settings differences in this instance cannot be attributed to
Rhoads and Morse (1971) and Savrda et al. (1984) differences in environmental parameters that can be
demonstrated that as the oxygen concentration in bottom assessed from the rock record (such as relative oxygen-
waters decreases so does the size of marine animals ation), but must be due to other factors that are more
inhabiting the substrate. In fossil and modern examples difficult to measure/observe from geochemical and facies
142 R.J. Twitchett / Palaeogeography, Palaeoclimatology, Palaeoecology 252 (2007) 132–144

and population. The latter two alternatives involve


decreases in population, which will increase the possibil-
ity of extinction as well as decrease the chances of
fossilisation (see also above discussion). Therefore, the
only taxa expected to be present in the fossil record (i.e.
actually found as fossils) during intervals of low primary
productivity and food shortage are those that have
undergone a reduction in body size.
Thus, food shortage could potentially be the cause of
within-lineage size decrease and temporary absence of
larger taxa from the fossil record (Figs. 6 and 7). It may
also be a reason for the extinction of large taxa (Fig. 4), as
these tend to have higher food requirements and lower
population sizes, which would make them vulnerable to
food shortage. Finally, if new taxa appeared during the
episode of low food supply, then they too would suffer the
same ecological constraints as the extinction survivors.
This could also explain why taxa appearing in the
Fig. 8. Size distributions of Claraia specimens from oxygen-restricted extinction aftermath were small (Fig. 5). Subsequent size
biofacies of the earliest Triassic from East Greenland (A) and Western increase would only take place as food supply increased,
Australia (B). East Greenland specimens were measured in the as “normal” primary production was resumed.
laboratory from samples collected in the field. Australian samples were A critical test for the food shortage hypothesis would be
measured from the Hovea-3 core.
to demonstrate that the Lilliput changes were synchronous
with a decrease in primary production. Good proxies for
studies (such as temperature, salinity, food supply, etc.). marine primary productivity exist in post-Jurassic strata,
Further comparative studies are clearly warranted. for example by comparing the δ13C values of benthic and
planktonic foraminifera (e.g. Zachos et al., 1989). In pre-
4.2.2. Food shortage Jurassic rocks, proxies for marine productivity are more
Food shortage is one environmental factor that has problematic and the δ13C record can be the result of
long been associated with body size decrease in animals several different processes such as terrestrial biomass
(Hallam, 1965), although in unicellular organisms, such burning, methane release, or volcanic activity (see Berner,
as benthic foraminifera, oligotrophic conditions often 2002 for recent discussion). When tackling these older
promote size increase as species seek to compensate for events, testing and rejecting alternative possible causes of
food shortage by incorporating symbionts (Brasier, 1995). size change (such as temperature change and benthic
Within-lineage size decrease (Fig. 7) has been linked to oxygen levels) will be one way forward, as the proxies for
productivity collapse and food shortage in Late Devonian these environmental changes are less equivocal than are
conodont taxa (Girard and Renaud, 1996). Many modern those for productivity levels. Some circumstantial evi-
studies show that growth rate and adult size in marine dence exists that the differences in the sizes of Claraia
animals are dependent upon, and correlated with, food from East Greenland and Australia (Fig. 8) may have been
supply (e.g. Kingsford and Hughes, 2005). In addition, caused by differences in food supply. Surface productivity
Twitchett (2001) recently proposed a model that links may indeed have been higher in Western Australia,
both within-lineage size decrease (Fig. 7) and temporary because sediments of the Hovea Member have a very
absence from the fossil record (Fig. 6) to an episode of high total organic carbon (TOC) content, and climate
primary productivity collapse and food shortage. If food models have indicated a major upwelling centre close by
supply is severely curtailed, then there will be an (Grice et al., 2005 and refs therein). More work is clearly
inevitable reduction in the biomass of taxa higher up the needed to resolve this issue.
food chain. Biomass is a function of both body size and
number of individual organisms. Therefore, the necessary 5. Summary
reduction can take place in one of three ways: (1) a
decrease in body size while maintaining population size; Qualitative observations concerning size change
(2) a decrease in the number of individuals while through the Permian–Triassic interval are being supple-
maintaining body size; (3) a decrease in both body size mented by more quantitative analyses. These demonstrate
R.J. Twitchett / Palaeogeography, Palaeoclimatology, Palaeoecology 252 (2007) 132–144 143

that all animal groups suffered a size reduction after the Erwin, D.H., 1993. The great Paleozoic crisis: life and death in the
Late Permian extinction event and that body sizes Permian. Columbia University Press, New York. 327 pp.
Erwin, D.H., 1994. The Permo-Triassic extinction. Nature 367, 231–236.
remained low, relative to earlier Permian or later Triassic Erwin, D.H., 1996. Understanding biotic recoveries: extinction, survival,
times, for the duration of the Early Triassic. In addition to and preservation during the end-Permian Mass Extinction. In:
this longer-term depression of body sizes, a shorter Jablonski, D., Erwin, D.H., Lipps, J.H. (Eds.), Evolutionary
duration, more severe body size reduction is observed in Paleobiology. University of Chicago Press, Chicago, pp. 398–418.
Fara, E., 2001. What are Lazarus taxa? Geological Journal 36, 291–303.
the immediate aftermath of the extinction event affecting
Flessa, K.W., Jablonski, D., 1983. Extinction is here to stay.
both survivors (i.e. the Lilliput effect) and newly Paleobiology 9, 315–321.
originating taxa. This phenomenon is only observed in Fraiser, M.L., Bottjer, D.J., 2004. The non-actualistic Early Triassic
high resolution data, and spans the parvus and isarcica gastropod fauna: a case study of the Lower Triassic Sinbad
zones of the Griesbachian (early Induan). From the cari- Limestone Member. Palaios 19, 259–275.
nata Zone onwards body size recovers somewhat, but Fraiser, M.L., Twitchett, R.J., Bottjer, D.J., 2005. Unique microgastropod
biofacies in the Early Triassic: indicator of long-term biotic stress and
remains lower than expected until the Middle Triassic. A the pattern of biotic recovery after the end-Permian mass extinction.
number of potential causes of the observed size changes Comptes Rendus Palevol 4, 543–552.
have been proposed and these can be tested from Girard, C., Renaud, S., 1996. Size variations in conodonts in response to
geological data. It is likely that a combination of factors the upper Kellwasser crisis (upper Devonian of the Montagne Noire,
was involved. Environmental parameters such as low France). Comptes Rendus de l'Academie des Sciences, Serie IIa 323,
435–442.
atmospheric and dissolved oxygen concentrations and Graham, J.B., Dudley, R., Aguilar, N.M., Gans, C., 1995. Implications of
food shortage caused by primary productivity collapse the Late Palaeozoic oxygen pulse for physiology and evolution.
may all have played a role. Nature 375, 117–120.
Grice, K., Cao, C., Love, G.D., Böttcher, M.E., Twitchett, R.J.,
Acknowledgements Grosjean, E., Summons, R.E., Turgeon, S.C., William Dunning,
W., Jin, Y., 2005. Photic zone euxinia during the Permian–
Triassic Superanoxic Event. Science 307, 706–709.
Caroline Hensley (Natural History Museum, London) is Hallam, A., 1965. Environmental causes of stunting in living and fossil
thanked for her help in accessing specimens in her care. marine benthonic invertebrates. Palaeontology 8, 132–155.
Rob Willink (Origin Energy) is thanked for inviting me to Hallam, A., 1975. Evolutionary size increase and longevity in Jurassic
study the Permian-Triassic Hovea-3 core in Perth. I have bivalves and ammonites. Nature 258, 493–496.
Harries, P.J., Kauffman, E.G., Hansen, T.A., 1996. Models for biotic
received encouragement and useful input over the years survival following mass extinction. In: Hart, M.B. (Ed.), Biotic
from numerous people, including Margaret Fraiser and Recovery from Mass Extinction Events. Geological Society of
Dave Bottjer (USC), and members of the Palaeo Discussion London Special Publications, vol. 102, pp. 41–60.
Group at Bristol (2002), particularly Mike Benton, Paul Harrison, J.F., Lighton, J.R.B., 1998. Oxygen-sensitive flight
metabolism in the dragonfly Erythemis simplicicollis. Journal of
Pearson (now Cardiff), and former students James Wheeley
Experimental Biology 201, 1739–1744.
and Nathan Price-Lloyd. P. Wignall and an anonymous Hayami, I., 1997. Size changes of bivalves and a hypothesis about the
reviewer are thanked for comments on an earlier draft. cause of mass extinction. Fossils 62, 24–36 (in Japanese).
Hayami, I., 1998. Ecology of mass extinctions: the diversity and shell
size of bivalves through time. Iden 52, 38–44 (in Japanese).
References
Hotinski, R.M., Bice, K.L., Kump, L.R., Najjar, R.G., Arthur, M.A.,
2001. Ocean stagnation and end-Permian anoxia. Geology 29, 7–10.
Barbault, R., 1988. Body size, ecological constraints and the evolution Isozaki, Y., 1997. Permo-Triassic boundary superanoxia and stratified
of life-history strategies. Evolutionary Biology 22, 261–286. superocean: records from lost deep sea. Science 276, 235–238.
Benton, M.J., 1995. Diversification and extinction in the history of life. Jablonski, D., 1996. Body size and macroevolution. In: Jablonski, D.,
Science 268, 52–58. Erwin, D.H., Lipps, J.H. (Eds.), Evolutionary Paleobiology. Chicago
Berner, R.A., 2002. Examination of hypotheses for the Permo-Triassic University Press, Chicago, pp. 256–289.
boundary extinction by carbon cycle modelling. Proceedings of the Jablonski, D., Raup, D.M., 1995. Selectivity of end-Cretaceous marine
National Academy of Sciences 99, 4172–4177. bivalve extinctions. Science 268, 389–391.
Berner, R.A., 2005. The carbon and sulfur cycles and atmospheric Jeffery, C.H., 2001. Heart urchins at the Cretaceous/Tertiary boundary:
oxygen from Middle Permian to Middle Triassic. Geochimica et a tale of two clades. Paleobiology 27, 140–158.
Cosmochimica Acta 69, 3211–3217. Kaljo, D., 1996. Diachronous recovery patterns in Early Silurian corals,
Brasier, M.D., 1995. Fossil indicators of nutrient levels. 2: evolution and graptolites and acritarchs. In: Hart, M.B. (Ed.), Biotic Recovery from
extinction in relation to oligotrophy. In: Bosence, D.W.J., Allison, P.A. Mass Extinction Events. Geological Society, London, Special
(Eds.), Marine Palaeoenvironmental Analysis From Fossils. Geolog- Publications, vol. 102, pp. 127–134.
ical Society Special Publication, vol. 83, pp. 133–150. Kauffman, E.G., Erwin, D.H., 1995. Surviving mass extinctions.
Cotgreave, P., 1993. The relationship between body size and Geotimes 14, 14–17.
population abundance in animals. Trends in Ecology and Kauffman, E.G., Harries, P.J., 1996. The importance of crisis progenitors
Evolution 8, 244–248. in recovery from mass extinction. In: Hart, M.B. (Ed.), Biotic
144 R.J. Twitchett / Palaeogeography, Palaeoclimatology, Palaeoecology 252 (2007) 132–144

Recovery from Mass Extinction Events. Geological Society, London, Schubert, J.K., Bottjer, D.J., 1995. Aftermath of the Permian–Triassic mass
Special Publications, vol. 102, pp. 15–39. extinction event: palaeoecology of Lower Triassic carbonates in the
Keller, G., 2003. Biotic effects of impacts and volcanism. Earth and western USA. Palaeogeography, Palaeoclimatology, Palaeoecology
Planetary Science Letters 215, 249–264. 116, 1–39.
Kingsford, M.J., Hughes, J.M., 2005. Patterns of growth, mortality and Smith, R.M.H., 1995. Changing fluvial environments across the
size of the tropical damselfish Acanthochromis polacanthus across Permian–Triassic boundary in the Karoo Basin, South Africa, and
the continental shelf of the Great Barrier Reef. Fishery Bulletin possible causes of tetrapod extinctions. Palaeogeography, Palaeo-
103, 561–573. climatology, Palaeoecology 117, 81–104.
Krystyn, L., Baud, A., Richoz, S., Twitchett, R.J., 2003. A unique Smith, A.B., 2001. Large-scale heterogeneity of the fossil record:
Permian–Triassic boundary section from Oman. Palaeogeography, implications for Phanerozoic biodiversity studies. Philosophical
Palaeoclimatology, Palaeoecology 191, 329–344. Transactions of the Royal Society of London. Series B 356, 351–367.
Labarbera, M., 1986. The evolution and ecology of body size. In: Smith, A.B., Jeffrey, C.H., 1998. Selectivity of extinction among sea
Raup, D.M., Jablonski, D. (Eds.), Patterns and Processes in the urchins at the end of the Cretaceous period. Nature 392, 69–71.
History of Life. Springer-Verlag, Berlin, pp. 69–98. Stanley, S.M., 1973. An explanation for Cope's Rule. Evolution 27,
Lockwood, R., 2005. Body size, extinction events, and the early 1–26.
Cenozoic record of veneroid bivalves: a new role for recoveries? Tverdokhlebov, V.P., Tverdokhlebov, G.I., Surkov, M.V., Benton, M.J.,
Paleobiology 31, 578–590. 2002. Tetrapod localities from the Triassic of the SE of European
Marshall, C.R., 1998. Determining stratigraphic ranges. In: Donovan, Russia. Earth Science Reviews 60, 1–66.
S.K., Paul, C.R.C. (Eds.), The Adequacy of the Fossil Record. Twitchett, R.J., 1999. Palaeoenvironments and faunal recovery after the
John Wiley and Sons, Chichester, pp. 23–54. end-Permian mass extinction. Palaeogeography, Palaeoclimatology,
McGhee, G.R., Sheehan, P.M., Bottjer, D.J., Droser, M.L., 2004. Palaeoecology 154, 27–37.
Ecological ranking of Phanerozoic biodiversity crises: ecological Twitchett, R.J., 2001. Incompleteness of the Permian–Triassic fossil
and taxonomic severities are decoupled. Palaeogeography, Palaeo- record: a consequence of productivity decline? Geological Journal
climatology, Palaeoecology 211, 289–297. 36, 341–353.
Newell, N.D., 1952. Periodicity in invertebrate evolution. Journal of Twitchett, R.J., 2006. The palaeoclimatology, palaeoecology and
Paleontology 26, 371–385. palaeoenvironmental analysis of mass extinction events. Palaeo-
Paul, C.R.C., Donovan, S.K., 1998. An overview of the completeness of geography, Palaeoclimatology, Palaeoecology 232, 190–213.
the fossil record. In: Donovan, S.K., Paul, C.R.C. (Eds.), The Adequacy Twitchett, R.J., Krystyn, L., Baud, A., Wheeley, J.R., Richoz, S., 2004.
of the Fossil Record. John Wiley and Sons, Chichester, pp. 111–132. Rapid marine recovery after the end-Permian mass extinction event
Payne, J.L., 2005. Evolutionary dynamics of gastropod size across the in the absence of marine anoxia. Geology 32, 805–808.
end-Permian extinction and through the Triassic recovery interval. Twitchett, R.J., Feinberg, J.M., O'Connor, D.D., Alvarez, W., McCollum, L.,
Paleobiology 31, 269–290. 2005. Early Triassic Ophiuroids: their paleoecology, taphonomy and
Price-Lloyd, N., Twitchett, R.J., 2002. The Lilliput effect in the aftermath distribution. Palaios 20, 213–223.
of the end-Permian mass extinction event. GSA Abstracts with Urbanek, A., 1993. Biotic crises in the history of Upper Silurian
Programs 34 (6), 355. graptoloids: a palaeobiological model. Historical Biology 7, 29–50.
Renaud, S., Girard, C., 1999. Strategies of survival during extreme Wheeley, J.R., Twitchett, R.J., 2005. Palaeoecological significance of
environmental perturbations: evolution of conodonts in response to a new Griesbachian (Early Triassic) gastropod assemblage from
the Kellwasser crisis (Upper Devonian). Palaeogeography, Palaeo- Oman. Lethaia 38, 37–45.
climatology, Palaeoecology 146, 19–32. Wignall, P.B., Benton, M.J., 1999. Lazarus taxa and fossil abundance
Rhoads, D.C., Morse, J.W., 1971. Evolutionary and ecologic significance at times of biotic crisis. Journal of the Geological Society of
of oxygen-deficient marine basins. Lethaia 4, 413–428. London 156, 453–456.
Richmond, C., Marcus, N.H., Sedlacek, C., Miller, G.A., Oppert, C., Wignall, P.B., Twitchett, R.J., 1996. Oceanic anoxia and the end
2006. Hypoxia and seasonal temperature: short-term effects and long Permian mass extinction. Science 272, 1155–1158.
term implications for Acartia tonsa dana. Journal of Experimental Wignall, P.B., Twitchett, R.J., 2002. Extent, duration and nature of the
Marine Biology and Ecology 328, 177–196. Permian–Triassic superanoxic event. In: Koeberl, C., MacLeod, K.G.
Savrda, C.E., Bottjer, D.J., 1986. Trace-fossil model for reconstruction (Eds.), Catastrophic Events and Mass Extinctions: Impacts and
of paleo-oxygenation in bottom waters. Geology 14, 3–6. Beyond. Geological Society of America Special Paper, vol. 356,
Savrda, C.E., Bottjer, D.J., Gorsline, D.S., 1984. Development of a pp. 395–413.
comprehensive oxygen-deficient marine biofacies model: evidence Zachos, J.C., Arthur, M.A., Dean, W.E., 1989. Geochemical evidence
from Santa Monica, San Pedro and Santa Barbara Basins, California for suppression of pelagic marine productivity at the Cretaceous–
continental borderland. American Association of Petroleum Geolo- Tertiary boundary. Nature 337, 61–64.
gists Bulletin 68, 1179–1192.

You might also like