The Eocene-Oligocene Transition

Abstract

A diverse array of fossil, geochemical and sedimentary data shows patterns of major change at or near the Eocene-Oligocene boundary, indicating a period of fundamental climatic and biotic reorganization on Earth. Multiple lines of evidence support the hypothesis that these changes signal major global cooling, especially at high latitudes, and rapid growth of semi-permanent ice-sheet on Antarctica in the early Oligocene. The quality and temporal resolution of Eocene-Oligocene fossil and sediment records and the diversity of climatic proxy tools have increased enormously in recent years, bringing a new level of detail to the study of this transition. The higher-resolution records have revealed that the climatic shifts across the Eocene-Oligocene boundary occurred in multiple stages that can be calibrated on orbital timescales. Coupled with incresasingly sophisticated computer models, these records have led to significant advances in our understanding of possible causal mechanisms and feedbacks driving this global shift. Here we review the wealth of evidence for Eocene-Oligocene climate change, summarize the current state of understanding, and highlight the key areas still requiring work that might guide the direction of future research. Obtaining a thorough understanding of this critical climatic transition is important for highlighting the mechanisms and sensitivities of Cenozoic climate, and addressing topical questions relating to the dynamics of global change during greenhouse-icehouse climate switching.

Full text

The Eocene–Oligocene Transition
H. K. COXALL & P. N. PEARSON
School of Earth, Ocean and Planetary Sciences, Cardiff University, Main Building, Park Place,
Cardiff, CF10 3AT, UK (e-mail: Helen.Coxall@earth.cf.ac.uk)
Abstract: A diverse array of fossil, geochemical and sedimentary data shows patterns of major
change at or near the Eocene –Oligocene boundary, indicating a period of fundamental climatic
and biotic reorganization on Earth. Multiple lines of evidence support the hypothesis that these
changes signal major global cooling, especially at high latitudes, and rapid growth of semi-
permanent ice-sheet on Antarctica in the early Oligocene. The quality and temporal resolution
of Eocene–Oligocene fossil and sediment records and the diversity of climatic proxy tools
have increased enormously in recent years, bringing a new level of detail to the study of this
transition. The higher-resolution records have revealed that the climatic shifts across the
Eocene– Oligocene boundary occurred in multiple stages that can be calibrated on orbital time-
scales. Coupled with incresasingly sophisticated computer models, these records have led to
significant advances in our understanding of possible causal mechanisms and feedbacks driving
this global shift. Here we review the wealth of evidence for Eocene –Oligocene climate change,
summarize the current state of understanding, and highlight the key areas still requiring work
that might guide the direction of future research. Obtaining a thorough understanding of this
critical climatic transition is important for highlighting the mechanisms and sensitivities of
Cenozoic climate, and addressing topical questions relating to the dynamics of global change
during greenhouse– icehouse climate switching.
The transition from the Eocene to the Oligocene
was a period of global change lasting about
500,000 years that marks a major step towards the
development of the modern glaciated climate.
Several decades of research (reviewed in this
paper), which builds on work stretching back to
the early 20th century, have revealed that this inter-
val is associated with extinctions and evolutionary
turnover, on land and in the oceans, and major
shifts in geochemical and sedimentological
proxies. These records provide strong evidence for
a phase of oceanic reorganization, global cooling
and the growth of the first semi-permanent
continental-scale ice-sheets on Antarctica. The cau-
sative mechanisms of Eocene–Oligocene (E –O)
climate change are widely debated, with much
of the discussion centring on the relative roles of
declining greenhouse gases and the opening of
Southern Hemisphere oceanic gateways in permit-
ting substantial ice build-up on Antarctica. Follow-
ing Quaternary and Neogene models, attention is
now being focused on defining the significance of
orbital configurations, which affect the distribution
of solar radiation received by the Earth, as well as
the role of many possible feedbacks within the
Earth System that may have interacted to give the
record of E–O change that we see.
Terrestrial and shallow marine records of E –O
climate change from around the world reveal sig-
nificant biotic turnover in plants and animals
across a range of latitudes that are linked to climatic
cooling, widespread regression and changes in the
hydrological cycle. These records are sporadic,
however, and the most complete archives are
found in the deep-sea realm where sedimentation
is often more continuous. Here, E –O environmental
changes have left their mark as major shifts in
microfossil communities, microfossil geochemis-
try, sedimentation and mineralogy. Much of the
recent progress in understanding the E –O transition
has come from palaeoclimate proxies derived from
these marine records that are accessed through
deep-sea coring.
A major problem has been that most Deep-Sea
Drilling Project (DSDP) and Ocean Drilling
Program (ODP) sites spanning the E –O boundary
are condensed and/or interrupted by hiatuses,
which have been attributed to an increase in ocean
circulation vigour and glacioeustatic sea-level fall
associated with climate change (Kennett &
Shackleton 1976; Aubry 1991; Miller et al. 1991;
Zachos et al. 1996). There has, therefore, been a
shortage of sequences appropriate for continuous
palaeoclimate analysis. Recent advances in drilling
technology and methods of stratigraphic correlation
have lead to improved recovery of deep-sea Palaeo-
gene sequences, and the quality of E –O sediment
archives and derived proxy records has increased
significantly. These newly available records,
combined with continuously more sophisticated
computer models, have led to advances in our
understanding of the timing and mechanisms of
From: WILL IAM S, M., HAY WOOD , A. M., GREG ORY , F. J. & SCHMI DT, D. N. (eds) Deep-Time Perspectives on Climate
Change: Marrying the Signal from Computer Models and Biological Proxies. The Micropalaeontological Society,
Special Publications. The Geological Society, London, 351 –387.
1747-602X/07/$15.00 #The Micropalaeontological Society 2007.

E–O climate change, and helped constrain theories
on the feedbacks involved in the inferred cryo-
sphere and biosphere responses.
Here we review the current state of understand-
ing of the E–O transition. The account will begin
with a discussion of the systematics of E –O bound-
ary stratigraphic terminology. The core of the paper
will review evidence for climatic change from (i)
fossils and (ii) palaeoclimatic proxies. Finally, we
provide a short synthesis summarizing progress to
date and our vision of where future research is
and should be directed.
Terminology, correlation and calibration
The E–O boundary is formally defined at the
Global Stratotype Section and Point (GSSP) at Mas-
signano, Italy, and corresponds to the extinction of
the planktonic foraminiferal Family Hantkeninidae
(see Coccioni et al. 1988; Nocchi et al. 1988;
Premoli Silva & Jenkins 1993; Berggren et al.
1995). The boundary is a tie-point in the timescale
of Cande & Kent (1995) where it is fixed at
33.7 Ma, although this figure is likely to be
refined by astronomical tuning (Coxall et al.
2005; Gale et al. 2006; Jovane et al. 2006). We
use the term ‘Eocene–Oligocene Transition’
(EOT) to encompass a phase of accelerated climatic
and biotic change lasting 500 kyr that began before
and ended after the boundary. This transition inter-
val is most clearly recognized using deep-sea
benthic foraminifera stable isotope data. In most
records, it begins with a phase of relatively negative
carbon isotopes and ends with a peak in both
oxygen and carbon isotope ratios. Our definition
of the transition interval excludes longer-term
events that have sometimes been included in discus-
sion of ‘terminal Eocene events’. We start with a
clarification and revision of the nomenclature.
The d
18
O peak at the end of the transition inter-
val coincides with the base of an Oligocene isotope
Zone defined by Miller et al. (1991) as ‘Oi-1’, and is
widely regarded as signalling a peak in glaciation
on Antarctica (Fig. 1). There is some confusion in
the literature regarding the meaning of Oi-1,
despite the fact that it was clearly defined by
Miller et al. (1991). Most workers use Oi-1 to
identify the climax of the early Oligocene d
18
O
excursion that lasted c. 400 kyr (Zachos et al.
1996; Coxall et al. 2005). As an isotope zone,
however, Oi-1 extends between two peaks in the
oxygen isotope record, and in fact spans much of
the lower Oligocene, equivalent to several million
years of time (Miller et al. 1991, Fig. 6, p. 6839).
The base of the Oi-1 zone is defined in DSDP Site
522 by the maximum d
18
O value in benthic forami-
nifera Stilostomella spp. (Oberhansli et al. 1984).
This occurs at 133.13 mbsf, which equates to
33.494 Ma on the geomagnetic polarity timescale
of Cande & Kent (1995) (19.38% through C13n)
(Miller et al. 1991; see summary by Vergnaud-
Grazzini & Oberhansli 1986). This level is slightly
above the level of the abrupt shift to high d
18
O
recorded in Cibicidoides spp. from the same site
(133.59 m, c.33.588 Ma), which represents the
base of ‘Oi-1’ as identified by Zachos et al.
(1996). Oi-1 does not correspond to or include an
isotopic shift, as implied in some of the literature
(e.g. Zachos et al. 2001, Fig. 2, p. 688; Gale et al.
2006, p. 412; Van Mourik & Brinkhuis 2005, p. 13).
The isotopic shift and the d
18
O maximum are
clearly both important events that represent differ-
ent climatic and environmental phases. Here we
use the term ‘Early Oligocene Glacial Maximum’
(EOGM), advocated by Liu et al. (2004) and Tuo
et al. (2006) (after Zachos et al. 1996) to identify
the phase of maximum d
18
O (Oi-1 of Zachos
et al. 1996), and to differentiate from Miller et al.’s
(1991) isotope zone. The phase of rapidly increas-
ing d
18
O that precedes the EOGM we refer to as
the E-O ‘Shift’. In the type section (DSDP Site
522) and elsewhere (Zachos et al. 1996; Coxall
et al. 2005), the Shift occurs over a period of
several hundred thousand years and encompasses
the E-O boundary. In some isotope records (see
below), it is resolved as two or more steps
(Fig. 1). The ‘E-O Transition’ in its strict sense,
therefore, is equivalent to the ‘E-O Shift’. Under
an alternative astronomical naming scheme (based
on the 400 kyr cycle of Earth’s eccentricity) that
has recently been proposed for the Oligocene gla-
ciations (Wade & Palike 2005) the EOGM is
referred to as event ‘84Eo C13n’.
Fossils
The fossil record across the E –O boundary has
been periodically reviewed as knowledge has
increased (Pomerol & Premoli Silva 1986;
Prothero & Berggren 1992; Prothero 1994; Ivany
et al. 2003; Prothero et al. 2003) and new data con-
tinue to be published every year. While the E –O
transition interval was not one of the ‘Big Five
mass extinctions’ (Sepkoski 1986), it was never-
theless a time of substantial extinction and ecologi-
cal reorganization in many biological groups.
Some of these groups are vital for biostratigraphic
correlation, so the disappearance or appearance of
key taxa in the fossil record is often tied up with
attempts to recognize the E–O boundary itself in
different facies. Because of fundamental problems
in precisely correlating stratigraphic sections
across environments, oceans, continents and lati-
tudes, the detailed sequence of biotic events is
H. K. COXALL & P. N. PEARSON352

still not clear at high resolution (Ivany et al. 2003).
However, it is likely that some of the extinctions
are associated with the phases of rapid climatic
change and sea-level fall across the E –O boundary
and others may be associated with the maximum
glacial conditions of the early Oligocene
(EOGM, Fig. 1; corresponding to the base of the
‘Oi-1’ isotope zone of Miller et al. 1991). The
radiation of more cold-adapted forms in many
biotic groups probably began during this glacial
period, and continued thereafter into the Oligo-
cene. Ongoing efforts directed at relating patterns
of biotic turnover in different environments to the
global isotope and palaeomagnetic record is
likely to lead to greater clarity as work progresses.
Planktonic foraminifera
In the Global Stratotype Section at Massignano in
Italy, the E–O boundary marker (the ‘Golden
Spike’) is placed at the last occurrence of the plank-
tonic foraminiferal Family Hantkeninidae (Nocchi
et al. 1988; Premoli Silva & Jenkins 1993). This
Family, although seldom dominant in planktonic
Fig. 1. Terminology, calibration and d
18
O signature of the Eocene–Oligocene Transition. (a) Benthic
foraminifera d
18
O compilation (after Zachos et al. 2001) from mid- to high southern latitudes (diamonds) and the
equatorial Pacific (other symbols) plotted on a common time scale (Berggren et al. 1995) showing the globally
recognizable c.d
18
O 1.5‰ shift and positive d
18
O anomaly. The extinction of Hantkenina spp. has been identified in
South Atlantic DSDP Site 522 (136.7 mbsf; Poore 1984) and is here calibrated (arrow) against the corresponding
benthic d
18
O record (black diamonds). (b)Cibicidoides spp. d
18
O (þ0.64) from ODP Site 1218 (3800 m water depth;
Coxall et al. 2005) (5-point moving average trend line) plotted on an orbitally tuned timescale. We distinguish the
interval of d
18
O ‘Shift’ (which is time equivalent to the Eocene-Oligocene Transition), from the early Oligocene d
18
O
maximum, here referred to as the Early Oligocene Glacial Maximum (EOGM) (after Zachos et al. 1996; Liu et al.
2004). These features are also recognizable in the compilation. The site 1218 data gap corresponds with a zone of
severe dissolution through which no calcareous benthic forams were available (see below). Lower d
18
O values at Site
1218 and other tropical sites suggest bottom-water temperatures c. 2 8C warmer than high southern latitudes.
Species-specific adjustments have been applied (þ0.64 or þ0.4) to account for vital effects (after Zachos et al. 2001).
THE E–O TRANSITION 353

assemblages, is a very distinctive component of
middle and upper Eocene pelagic carbonates world-
wide (Fig. 2). Coccioni et al. (1988) recognized five
species and two genera (Hantkenina and Cribro-
hantkenina) of Hantkeninidae at Massignano,
although in their analysis they do not all persist to
the boundary, and a similar set of species can be
recognized in the Spanish sections such as that at
Fuente Caldera (Molina 1986; Molina et al.
2006). The earlier suggestion by Blow (1979) that
the extinction of Cribrohantkenina preceded
Hantkenina (which was based on limited sampling)
can now be discounted. Our own work (Coxall &
Pearson 2006) suggests that the extinction of
Hantkenina and Cribrohantkenina was essentially
simultaneous and involved all five species. More-
over, it exactly coincides with the local extinction
of another common species, Pseudohastigerina
micra (Molina et al. 2006), or its dwarfing (Nocchi
et al. 1986), and is close to the first appearance of
a typically Oligocene form, Globoquadrina tapur-
iensis (Blow & Banner 1962; Coccioni et al. 1988;
see also Keller 1983 and Keller et al. 1992). This
means that the E–O boundary is one of the best-
defined biostratigraphic levels of the Cenozoic for
planktonic foraminifera and in the latest tropical
subtropical biozonation it is used as the top of the
topmost Eocene zone, E16 (Berggren & Pearson
2005). Although good stratigraphic sections with
acceptable carbonate preservation are rare, there is
as yet no evidence that the extinction of the
Hantkeninidae was locally controlled or diachro-
nous (as claimed by Van Mourick & Brinkhuis
2005), except in the high polar latitudes where the
group only occasionally occurs during climatically
favourable episodes.
Another major extinction in the planktonic fora-
miniferal records can be found close to the last
occurrence of the Hantkeninidae, namely the
extinction of the Turborotalia cerroazulensis
group, which in most taxonomies consist of three
separate species (see Pearson et al. 2006 for a
recent review). In the Massignano stratotype, this
occurs just 60 cm below the Golden Spike, which
equates to about 65 kyr (Coccioni et al. 1988;
Berggren & Pearson 2005). It also narrowly pre-
dates the Hantkenina extinction at Fuente Caldera
in Spain (Molina et al. 2006) and several deep-sea
drilling sites in the North and South Atlantic and
Indian Oceans (e.g. Poore 1984; Pearson &
Chaisson 1997). The T. cerroazulensis group of
species was very abundant and widespread. Both
the Hantkeninidae and the T. cerroazulensis group
existed for many millions of years before their
eventual demise, which hints that their closely
spaced extinctions were very likely more than
coincidental and related to a prolonged phase of
environmental disruption. Taken together, they rep-
resent one of the most obvious extinctions of the
Cenozoic among planktonic foraminifera.
Boersma & Premoli Silva (1986) and Keller
et al. (1992) noted that the long-term trend of
planktonic foraminifera evolutionary turnover
from the middle Eocene into the Oligocene
largely involves extinction of warm-water, tropical,
surface-dwelling species. The extinctions at and
near the E–O boundary might therefore be due to
rapid environmental change and cooling, and may
also have been influenced by changing water mass
stratification and patterns of biological productivity.
Unfortunately, it still not possible to confidently
(and precisely) locate the planktonic foraminifer
Fig. 2. SEM micrographs of species of the planktonic foraminiferal family Hantkeninidae that went extinct at the
E–O boundary (see Coxall & Pearson 2006). (1)Hantkenina alabamensis; (2)H. primitiva; (3)H. compressa;
(4)H. nanggulanensis; and (5)Cribrohantkenina inflata. Specimens are from various deep-sea sites. Scale
bar ¼100 mm. The extinction of these species denotes the Eocene –Oligocene boundary worldwide.
H. K. COXALL & P. N. PEARSON354

extinctions with reference to the stable isotope
events (see below). Probably the best published
record is DSDP Site 522 on the Walvis Ridge,
South Atlantic, where the hantkeninid extinction
(Poore 1984) occurs within the isotope shift that
precedes the basal EOGM (see Fig. 1), but in Site
522 the planktonic foraminifera are somewhat dis-
solved and fragmentary. Another relevant section
is ODP Site 925 in the tropical North Atlantic.
Here the preservation is even worse (in hard lime-
stone), but the hantkeninid extinction (Pearson &
Chaisson 1997) occurs above a zone of dissolution
and below the most prominent carbon isotopic shift
that leads into the EOGM (Diester-Haass & Zachos
2003).
Nannofossils
The other calcareous group that is widely used for
deep-sea biostratigraphy is the nannoplankton.
Perch-Nielsen (1986) and Aubry (1992) documen-
ted a long-term decline in diversity from the
middle Eocene into the Oligocene but little
change is associated with the boundary interval
itself. The only significant extinctions that occur
near the E–O boundary interval are the stepped dis-
appearance of rosette-shaped discoasters (Aubry
1992). The disappearance of Discoaster saipanensis
at approximately 34.2 Ma (i.e. about 500 kyr before
the boundary: Berggren et al. 1995) is preceded
slightly by the extinction of another species, D. bar-
badiensis. The former of these extinctions marks
the boundary between nannofossil Zones NP20
and NP21. A more minor event, namely the extinc-
tion of Pemma papillatum, may be more closely
associated with the E–O boundary itself (Varol
1998) but requires further study. Beyond that, no
major changes to nannofossil assemblages have
been reported, but there are significant biogeographic
changes associated with the basal Oligocene glacia-
tion and the onset of a relatively unstratified ocean
around Antarctica, where nannoplankton locally dis-
appear from the record (Hay et al. 2005).
Radiolaria
Although it has long been thought that radiolaria
were largely unaffected by the E –O transition
(e.g. Riedel & Sanfillipo 1986), a major turnover
of tropical species has recently been described by
Funakawa et al. (2006) from sites drilled in the
Pacific. This turnover includes the extinction of
several taxa and a corresponding sudden drop in
diversity and radiolarian accumulation rates, and
is combined with the expansion of cool-water cos-
mopolitan taxa. It appears to be closely associated
with the stable isotope shifts that precede the
EOGM and awaits study in other areas, especially
high latitudes.
Dinoflagellate cysts
Brinkhuis (1992), Brinkhuis & Biffi (1993) and
more recently Van Mourik & Brinkhuis (2005)
have reviewed the evolution of dinoflagellate
cysts (dinocysts) across the Eocene –Oligocene
boundary interval in Italy. Van Mourik & Brinkhuis
(2005) present new data from the GSSP and a
neighbouring sediment core (the ‘Massicore’).
Across the E–O boundary interval in Italy, two suc-
cessive influxes of cool-water high-latitude species
occur (Brinkhuis & Biffi 1993), the first of which
correlates directly with the E–O boundary sensu
stricto and the second with the onset of a more
severe cold episode and inferred sea-level lowstand.
Although these effects are local migrations, they
may relate to the stepwise isotopic shifts that
precede the early Oligocene glaciation and the
maximum glaciation itself. A later event, identified
in the Massicore, is the extinction of Areo-
sphaeridium diktyoplokum, which occurs near
the top of Chron C13n. This is several hundred
thousand years younger than the d
18
O maximum.
Elsewhere, the record of dinocysts across the
E–O boundary interval is patchy with the record
from local basins probably affected by sea-level
fall and changing local environments (e.g. Gedl
2004). Sluijs et al. (2003) have described the tran-
sition interval in the Southern Ocean where typical
early Palaeogene assemblages are sequentially
replaced with assemblages dominated by Brigantedi-
nium, which is interpreted as related to the onset of
upwelling conditions in the water column. Records
from the Weddell Sea and off Dronning Maud
Land (Antarctic margin) show decreasing dinocyst
diversity through the Eocene and at the E–O bound-
ary, falling to only twospecies by the end of the early
Oligocene (Mohr 1990). This pattern is thought to
indicate development of cold surface waters.
Diatoms
Baldauf (1992) brought together deep-sea drilling
data available at that time and identified a signifi-
cant turnover in marine diatoms as occurring at
the E–O transition, with a notable increase in diver-
sity in the tropics. He interpreted this as being
related to a decrease in the vertical stratification
of the water column and an increase in the latitudi-
nal thermal gradient. The most profound changes
seem to have occurred in the higher latitudes,
associated with the changes in high-latitude water
masses across the E–O transition. For example,
Suto (2006) noted a rapid diversification of Chaeto-
ceros resting spores in the Norwegian Sea, which
THE E–O TRANSITION 355

they associate with a change from a stable water
column with a constant nutrient supply in the
Eocene to an unstable and more vertically mixed
water column in the Oligocene. Similarly in the
southern high latitudes, Olney et al. (2005) record
an increase in cold-adapted species in the Oligocene
of Antarctica, and Whitehead (2005) describes an
increase in the diversity of benthic diatoms in the
Oligocene associated with the increasing influence
of the Antarctic polar current.
Benthic foraminifera
Benthic foraminifera occur in a range of environ-
ments and habitats, and the pattern of evolution
differs markedly between these habitats. In
abyssal and bathyal environments, there is little
major change associated with the boundary interval
itself (Thomas 1992; Coccioni & Galeotti 2003).
Nevertheless, Diester-Haass & Zahn (2001) and
Diester-Haass & Zachos (2003) record a sudden
increase in the rate of accumulation of benthic
foraminifera in deep-sea environments in various
parts of the world, which they correlate to the
rapid isotopic shifts and attribute to an increase in
productivity associated with the more vigorous
overturning of the earliest Oligocene ocean (see
also Boersma 1986, for a similar suggestion).
Thomas and Gooday (1996) also observe declining
deep-sea benthic diversity and an increase in dom-
inance of opportunistic phytodetritus-exploiting
species (e.g. Epistominella exigua and Alabami-
nella weddellensis) that suggest a switch to a
more unpredictable and seasonally fluctuating
food supply, especially at high latitudes during the
transition. Kaminski (2005) has recorded an acme
of deep-water agglutinated species in the North
Atlantic and Western Tethys during the E –O tran-
sition interval, one of several such events in the
Palaeogene.
The pattern of turnover in smaller benthic fora-
minifera from shallower environments seems to
contrast with the deep-sea record in that significant
changes are more obvious. McGowran & Beecroft
(1986) and McGowran et al. (1992) noted an
abrupt turnover of neritic benthic foraminifera in
Australia, which they linked to the sudden cooling
and sea-level drop associated with the early Oligo-
cene glaciation. The pattern is similar in the US
Gulf Coast, although, as in Australia, it is compli-
cated by facies changes across the boundary
(Fluegeman 2003). A slightly delayed radiation of
Oligocene shallow-water smaller benthic foramini-
fera in the US Gulf Coast was noted by Fluegeman
(2003).
By far the most dramatic evidence for extinction
and turnover occurs in larger foraminifera. These
organisms were important carbonate producers
throughout the Eocene, where they were abundant
and widespread. According to Kiessling et al.
(2003) and Nebelsick et al. (2005), Eocene carbonate
platforms, which are dominated by coralline algae
and larger foraminifera, declined rapidly in the
Late Eocene and across the E–O boundary interval,
reaching a ‘post-Cambrian low’ in the earliest
Oligocene. Unfortunately, continuous sections in
carbonate facies are rare, possibly because of the sea-
level drops associated with the glacial maximum.
The most complete section available seems to be
the Melinau limestone of Sarawak, which was
studied in detail by Adams (1965). Adams et al.
(1986) reviewed the Indo-Pacific records and noted
a mass extinction of some important long-ranging
genera and species, notably Asterocyclina,Discocy-
clina and some species of Nummulites. These events
occur in the Melinau limestone associated with a
brief switch to algal-dominated facies. The mass
extinction has not been reliably correlated to
isotope or magnetostratigraphy, nor has it been
demonstrated how abrupt or gradual the extinctions
were, but Adams et al. (1986) suggest a relation to
the climate deterioration and sea-level fall across
the E–O transition. The Americas were a different
faunal province, but from limited evidence, the
pattern may be similar; for example, Robinson
(2003) has noted a rapid decline in larger foramini-
fera in the Caribbean that may be associated with
the boundary interval.
Ostracodes
Benson (1975) suggested that the E–O transition
to be the most significant in the Cenozoic history
of ostracodes, but it is not clear how precisely this
turnover correlates to the boundary events and
to what extent different habitats were affected.
Schellenberg (1998) found significant drops in diver-
sity of ostracodes across the E–O boundary interval,
and noted that the extinction seems to have been
most severe among deposit-feeding species. In con-
trast, Dall’Antonia et al. (2003) found little variation
in assemblages across the E– O boundary in the
Massignano stratotype section, and found little
change in either the shallow and deep records
from the US Gulf Coast and Barbados respectively,
except for some minor extinctions in both. Like
benthic foraminifera, ostracodes lived in a wide
variety of habitats and the Eocene turnover pattern
probably varies between them.
Shallow marine invertebrates
The fossil record of shallow marine invertebrates
across the E–O boundary worldwide is quite
patchy, with the best records coming from North
America and Europe. Dockery (1986) and
H. K. COXALL & P. N. PEARSON356

Dockery & Lozouet (2003) have described a major
turnover of molluscs in the early Oligocene of the
US Gulf Coast and Paris Basin, including the disap-
pearance of many long-ranging forms, which they
attribute to sea-level fall and global cooling.
Dockery & Lozouet (2003) also document a major
influx of European taxa into North America in the
earliest Oligocene that may be related to changing
water mass circulation in the North Atlantic.
Squires (2003), Hickman (2003) and Nesbitt
(2003) described a major and relatively abrupt turn-
over of marine molluscs in the earliest Oligocene in
the western United States, in which over half of pre-
existing genera disappeared. During these events,
warm-water taxa suffered disproportionately,
while cold-water forms radiated and expanded
their ranges. The higher-latitude record from
Alaska and Kamchatka (Oleinik & Marincovich
2003) is similar to the record further south. More
recently, minor extinction and size reduction in
veneroid bivalves has been described from the
eastern United States (Lockwood 2005). Other
events, both before and after the E–O boundary
interval, are also in the record and some groups,
while others such as the echinoids of the United
States (Carter 2003) do not show an unusual
extinction pattern.
The fundamental problem involved with
shallow-water marine invertebrates (as with
benthic foraminifera) is distinguishing global and
local effects. The problem of sea-level fall associ-
ated with the transition is bound to have caused
many local changes; far better geographic sampling
would be necessary to distinguish local from
global effects.
Terrestrial vegetation
There is clear evidence for major changes in terres-
trial vegetation worldwide across the E –O tran-
sition, but the pattern of change varied from
continent to continent and across the latitudes.
The change in North American floras was recog-
nized by Wolfe (1978) as the ‘Terminal Eocene
Event’ and has been studied in more detail by
Wolfe (1992, 1994) and recently by Liv et al.
(2007). The dominant pattern in North America is
for widespread replacement of the subtropical
broad-leaved Evergreen vegetation of the Eocene
with cooler, deciduous forms in the Oligocene,
with regional extinctions, especially at higher lati-
tudes. Similarly in the south, Patagonian floral
records show the disappearance of tropical veg-
etation and the rise to dominance of subtropical
and cool-temperate species from the middle
Eocene of Early Oligocene (Palazzessi & Barreda
2007). In other places in North America, there
was a diversification of desert species as aridity
increased in the continental interior (Yancey et al.
2003; Moore & Jansen 2005). In some areas such
as the US Gulf Coast, there is less evidence for dra-
matic change (Oboh-Ikuenobe & Jaramillo 2003).
Collinson (1992) has reviewed the botanical data
from Europe, where there is a similar change to a
more seasonal temperate flora accompanied by
loss of tropical and subtropical taxa. These
records suggest that the main phase of change
shortly pre-dates the mammalian extinctions
(discussed below) (see also Hooker et al. 2004).
The evolution of extensive North American
grassland habitats, and associated mammalian radi-
ation, has been linked with the Eocene –Oligocene
Transition but the timing is controversial (Stucky
1992; Stro
¨mberg 2005 and references therein). A
recent study suggests that the spread of this
habitat, and the mammalian grazing ecology,
occurred between the late Oligocene and Early
Miocene, driven by post-EOGM climate changes
that led to increased seasonal aridity (Stro
¨mberg
2005, 2006).
With regard to the tropics, Jaramillo et al. (2006)
have recently produced a detailed composite floral
record from Columbia and Venezuela, based on
pollen and spore occurrences that produce an unpre-
cedented view of long-term floral diversity. The
most rapid phase of change in this record occurs
around the E–O boundary interval, which Jaramillo
et al. (2006) link to the environmental changes
associated with the onset of Antarctic glaciation.
The resolution, however, is too low to correlate
floral changes with the events of the transition
itself, as seen in detailed deep-sea records (see
below).
Perhaps the most dramatic changes would be
expected in the high southern latitudes close to
the supposed expansion of Antarctic ice. Macro-
floral and palynological evidence from Antarctica
has been reviewed by Francis (1999) and Francis
& Poole (2002). The E –O transition interval
coincides with a change from evergreen forest to
sparse tundra on the Antarctic Peninsula and
Seymour Island, and Francis (1999) goes further
to suggest that vegetation–climate feedbacks may
have played a role in the rapid environmental
changes at that time. Similar patterns are seen in
the high northern latitudes, where there was wide-
spread displacement of subtropical broad-leaved
vegetation, especially at altitude (Wolfe 1994;
Myers 2003).
Evidence from beyond America and Europe is
sparser. Kemp (1978) reviewed the Australian
palaeobotanical record and found a major change
near the E–O boundary, with a trend towards
lower diversity, higher seasonality and a spread of
cool-temperate plants. Leopold et al. (1992) docu-
ment significant changes in Asian floras, with
THE E–O TRANSITION 357

diversity reduction across the boundary interval.
Ramussen et al. (1992) found little variation in
North Africa. More recently, however, Pan et al.
(2006) have suggested that the E–O transition
was the most important step in the extinction of
African palms (Arecaceae), although they contin-
ued to flourish on other continents.
Mammals
A mass extinction of mammalian faunas in Europe
(principally known from records in the Paris and
Hampshire basins) at approximately the E –O
boundary was noted almost a century ago by
Stehlin (1909) as the ‘Grand Coupure’ (great
break). It represents a major turnover in both peris-
sodactyls and artiodactyls (hoofed herbivores).
This has recently been dated as approximately
coinciding with the EOGM (Hooker et al. 2004)
and the associated sea-level regression. Approxi-
mately 60% of taxa seem to have disappeared at
this time and it has been argued that the extinctions
were caused either by climatic deterioration or com-
petition with immigrant taxa from Asia (Savage &
Russell 1983). A major turnover of mammals in
Mongolia at this time has also been linked to
climatic change (Meng & McKenna 1998).
In North America, the mammalian record has
long been confused because of a misalignment of
the land mammal stages with the global stratigraphy
(see Prothero & Swisher 1992 for discussion). It is
now clear, however, that there was a more minor
extinction, coinciding with the diversification of
hypsodont mammals close to the E –O boundary
(Stucky 1992), and there are similarities with the
record in South America (Marshall & Cifelli
1989). There may be a similar pattern in Asia but
the issue is further confused by uncertain chronol-
ogy (see comments by Berggren et al. 1992). The
Oligocene in general marks the time when represen-
tatives of modern land mammals became the
dominant vertebrate life, including the appearance
of the first primates and apes, but precise timing is
problematic.
In the marine realm, it has long been recognized
that the extinction of the archeocete whales such as
the giant Basilosaurus occurred at or near the E –O
transition (Fordyce 1992, 2003; Manning 2003)
although there is no evidence for a substantial
mass extinction (Fordyce 1992). The cold high-
latitude water masses of the early Oligocene seem
to have been a spur to the evolution of crown-group
whales, and the first baleen whales are known from
the early Oligocene (Fordyce 1992, 2003), most
likely in parallel with increased Southern Ocean
upwelling and increases in plankton abundance. In
addition, it has been demonstrated that toothed
whales (suborder Odontoceti), which include
dolphins and porpoises, underwent a significant
increase in brain size with respect to body size
near the E–O boundary (Marino et al. 2004).
Others
The amphibians and freshwater turtle records from
North America seem to suggest declining diversity
(Hutchinson 1992), possibly related to increasing
aridity around the E –O boundary. Corsini et al.
(2006) also identified changes in turtle carapace
size that may be a response to E –O climatic
change. A similar picture emerges from Europe
(Rage 1986) where tropical taxa disappear in large
numbers near the end of the Eocene. The first diver-
sification of cat-like carnivores (mostly sabre-
toothed) and other predators also occurs in the
early Oligocene and is thought to coincide with
the development of grassland habitats in central
North America (Bryant 1996). Evidence for other
tetrapods tends to be patchy and difficult to interpret
(e.g. see Prothero & Emry 1996), although it may be
noteworthy that there seems to have been a radi-
ation of passerine birds in the early Oligocene
(Mayr 2005).
Summary of palaeontological evidence
There is widespread evidence for enhanced extinc-
tion in many groups across the E –O boundary inter-
val, followed either immediately, or after a slight
delay, by renewed radiation in the Oligocene.
Precise correlations to the isotope shifts, glacial
maxima and magnetostratigraphy are still proble-
matic in many groups. In several groups where
there is sufficient age-discrimination, the extinc-
tions seem to have been stepwise over a period of
change that may equate to several hundred thousand
years, starting before the stage boundary as for-
mally defined and ending with the maximum glacia-
tion that corresponds to the base of the Oi-1 isotope
zone. It is interesting that in groups as diverse as for-
aminifera, primates, whales and birds, the Oligocene
saw the diversification of recognizably modern taxa,
whereas most of the Eocene forms are from ‘prehis-
toric’, now-extinct groups that lie outside of the
modern lineages. This pattern almost certainly
relates to the fact that the E–O boundary interval
was one of the most profound periods of change
that led to the modern climatic regime, the change
that is widely described as the greenhouse –icehouse
transition (Miller et al. 1991).
The close coincidence of widespread extinction
and stepwise climatic changes seems to imply
causal linkage, but the causes of extinction must
be different in many groups. Possibilities include
temperature change, sea-level fall and the associ-
ated exposure of the continental shelf, change in
H. K. COXALL & P. N. PEARSON358

biological production on land and in the oceans,
water mass properties (especially at high latitudes
but also in equatorial and upwelling areas), atmos-
pheric and ocean chemistry, etc. The whole Earth
System seems to have entered a period of prolonged
change, a factor that makes the E–O boundary
extinctions rather different from the more sudden
mass extinctions at the end of the Cretaceous and
probably at the end of the Permian as well. In the
following sections, we will review the geochemical
and other evidence for prolonged and interlinked
changes in the Earth System at this time.
Proxies
Other than the occurrence of glaciomarine sedi-
ments on and around Antarctica (e.g. Barrett et al.
1989; Zachos et al. 1999), there is little direct
evidence of E–O climate change. Therefore,
climate-sensitive proxies are used to trace responses
to climatic parameters indirectly. These proxies
provide information on global temperature, ice
volume, ocean productivity, water mass structure,
circulation, carbon cycling, atmospheric carbon
dioxide concentration and continental weathering.
Many of the processes being recorded are related to
the global carbon cycle, either as cause or effect,
and have the potential to influence and be influenced
by atmospheric CO
2
concentration. The search for
cause and effect relationship is therefore similar to
the lesser (but still-unresolved) problem of under-
standing glacial– interglacial climatic switching.
Other important techniques that are integral
to modern E–O palaeoclimatic study are climate
modelling on various scales and complexities, and
time series analysis, the latter having the power to
identify patterns of external orbital forcing on
Milankovitch timescales. Tectonic evolution is
also fundamental to the system and feeds into all
questions related to E–O climate change. Descrip-
tions and applications of the various proxies and
modelling techniques that have been utilized in the
study of the E–O transition are discussed below
under headings that represent the principal process
or parameter they are tracing. As with the discussion
of fossil evidence, there is no way it can be com-
pleted in this discussion, but we can identify a snap-
shot of important and promising areas.
Ice, temperature and sea level
Benthic foraminifera oxygen stable isotopes –
Undoubtedly, the most significant proxy for docu-
menting E–O climatic change is benthic forami-
nifera d
18
O. As classically demonstrated, the
principal controls on benthic d
18
O are: (i) seawater
temperature and, (ii) d
18
O of seawater (d
w
), which
changes during glaciations. The benthic d
18
O
proxy, therefore, helps constrain deep-water temp-
eratures and global ice volumes. Since the Southern
Ocean was probably the dominant deep-water
source from at least the Late Eocene (e.g. Wright
& Miller 1993), benthic d
18
O also provides an esti-
mate of high-latitude surface water temperatures.
As discussed above, studies of numerous
deep-sea cores demonstrates the existence of an
abrupt positive c. 1.2–1.5‰ shift in benthic d
18
O
associated with the E–O boundary that builds to a
sustained maximum at the base of the Oligocene
(Fig. 1). This interval of maximum d
18
O, which
we refer to as the EOGM, is now widely believed
to represent the maximum extent of the first semi-
permanent ice sheet on Antarctica (e.g. Kennett &
Shackleton 1976; Kennett 1977; Miller et al.
1987; Zachos et al. 1996, 1999, 2001; Lear et al.
2001; Coxall et al. 2005), 50% (or more) of the
present Antarctic ice-sheet volume (Barker et al.
1999; DeConto & Pollard 2003a). Following ter-
mination of the EOGM, d
18
O recovered to a new
equilibrium value that was on average 1‰ higher
than the Late Eocene, suggesting subsequent degla-
ciation and stabilization of Oligocene ice volumes.
High resolution (2– 10 kyr) d
18
O records have
improved constraints on the timing, magnitude
and structure of the E –O isotopic shift and d
18
O
maximum and identified additional structure
within it. These records demonstrate that the
EOGM can be correlated fairly precisely to the
base of magnetic Chron C13n with an estimated
duration of 400 kyr (e.g. Oberhansli et al. 1984;
Miller 1985; Zachos et al. 1996; Coxall et al.
2005). The c. 1.5‰ d
18
O shift into the EOGM
lasted about 500 kyr and occurred in two phases
(Zachos et al. 1996; Coxall et al. 2005). Orbitally
tuned records from equatorial Pacific ODP Site
1218 suggest that this involved two 40 kyr steps,
separated by a well-defined 200 kyr-long plateau
(Coxall et al. 2005) (Fig. 1b), although this is
not a unique interpretation of the orbital evidence.
The first step, which accounts for less than half
of the total shift, and the plateau occur in reversed
interval C13r and the second step was completed
at approximately the base of C13n. This relation-
ship to the magnetic reversal record holds up in
DSDP Site 522 and ODP site 744 (Zachos et al.
1996). The step-form is remarkably similar to the
pattern of non-linear ice growth simulated by a
coupled GCM-ice sheet models (DeConto &
Pollard 2003a,b), although it is of significantly
greater magnitude than predicted by the model. At
ODP Site 744 (Southern Indian Ocean), the
EOGM comprises two pronounced peaks lasting
approximately 100 and 150 kyr respectively.
These have been interpreted as two distinct
glacial maxima, termed Oi-1a and Oi-1b (Zachos
THE E–O TRANSITION 359

et al. 1996). These features are less well
defined elsewhere. The termination of the EOGM
is more difficult to define and correlate between
sites than its initiation because it consists of a series
of small stepped decreases in d
18
O, totalling c.
0.5– 0.6‰, interspersed with minor positive excur-
sions spread over a few tens of thousands of years
(Zachos et al. 1996; Coxall et al. 2005). Most of
this decrease, however, occurs in the top of magneto-
chron C13n (which is where the upper boundary of
the EOGM is placed), coincident with a decrease in
glaciomarine sediments close to Antarctica (e.g.
Wise et al. 1991).
Quantifying the relative contribution of temp-
erature and ice growth to the c. 1.5 ‰ d
18
O shift
is a problem, but separating the two effects is not
a trivial issue. One method is to obtain independent
palaeotemperature estimates using Mg/Ca ratios in
order to extract the d
w
component. This method is
summarized here but discussed fully by Lear
(2007). Other methods that have been used
include looking for covariance in deep-sea benthic
and low-latitude planktonic foraminifera d
18
O
records (e.g. Miller et al. 1991) and modelling of
Antarctic ice-sheet sensitivity to climate change
(Oerlemans 2004, 2005).
Magnesium calcium ratios and ice volume –
Benthic foraminiferal Mg/Ca palaeothermometry
has been useful for providing independent tempera-
ture estimates and, therefore, resolving temperature
and ice volume contributions to d
18
O records in the
Quaternary and early Cenozoic (Mashiotta et al.
1999; Lear et al. 2001, 2004; Martin et al. 2002;
Billups & Schrag 2003). Surprisingly, existing
E–O Mg/Ca records show no evidence of deep-
water cooling. In fact, if anything they suggest a
28C warming (Lear et al. 2001, 2004; Billups &
Schrag 2003) (Fig. 3a). Taken at face value, this
suggests that the entire E–O d
18
O shift is attribu-
table to ice growth. This implied ice volumes
seem unrealistic for Antarctica alone (Coxall et al.
2005), and raises several possibilities: (i) that
changes in the hydrological system (moisture
supply) rather than cool temperatures were import-
ant for ice-sheet expansion (Lear et al. 2001) (cf the
‘snow gun hypothesis’, Prentice & Matthews 1991),
(ii) there was more ice elsewhere; i.e. possible
contemporaneous northern hemisphere glaciation,
consistent with sedimentary and modelling
(DeCouto & Pollard 2007) evidence, (Davies
et al. 2001; Via & Thomas 2006; Eldrett et al.
2007), and/or (iii) that some additional factor acts
to mask the Mg/Ca cooling signal, such as seawater
pH and/or carbonate ion concentration (Billups &
Schrag 2003; Lear et al. 2004).
The lack of a cooling signal in deep water proxy
records is puzzling because, although cooling might
not have been the trigger for E –O Antarctica
glaciation, it seems logical that widespread Antarc-
tic glaciation would lead to local cooling in the
vicinity of Antarctica in the early Oligocene
(Zachos et al. 1996; Oerlemans 2005). A likely
explanation for the lack of Mg/Ca evidence for
cooling is that the extreme increase in deep ocean
carbonate ion concentration/alkalinity associated
with a 1 km deepening of the calcite compensation
depth (see below), synchronous with the shift into
the EOGM, masked the cooling (Lear et al. 2004;
Coxall et al. 2005). Until these effects are better
understood, the Mg/Ca record across the E–O
should be interpreted cautiously.
Covariance of d
18
O in global benthic and non-
upwelling tropical planktonic foraminifera (i.e.
from regions with the most thermally stable
surface waters) has been used as an indicator of
global ice volume fluctuations in the Quaternary
(e.g. Shackleton & Opdyke 1973). However, there
is a general shortage of suitable low-latitude and/
or high-resolution E–O planktonic records
because of poor preservation of dissolution-
susceptible planktonic shells in notoriously
carbonate-poor deep-sea sequences of this age.
Existing planktonic d
18
O records are mainly
restricted to the mid- to high latitudes, i.e. DSDP
Site 522 (Oberhansli et al. 1984), DSDP Sites 592
and 593 (Murphy & Kennett 1986) and ODP Site
744 (Stott et al. 1990; Barrera & Huber 1991).
These show highly variable increases in planktonic
d
18
O, ranging from 0.5 to 1.5 times benthic d
18
O,
that reflect complex contributions of regional sea-
surface temperatures, salinity, depth habitat and
post-burial diagenesis effects, in addition to sea-
level shift related to ice growth. Prentice &
Matthews (1988) compared compilations of low-
latitude planktonic and deep-sea benthic d
18
O
during the Cenozoic. They concluded that benthic
d
18
O primarily reflects deep-ocean temperature
variation whereas tropical planktonic d
18
O show
responses to Cenozoic ice volume and suggested
the presence of a significant ice budget for the
past 40 myr. The Prentice & Matthews (1988)
record, however, lacks the detail necessary to
resolve the pattern of change during the E–O tran-
sition. In any case, assumptions about the stability
of low-latitude temperatures during major climatic
transitions may be ill founded (Miller et al. 1991).
Acquisition of high-quality planktonic isotope
records from low and high latitudes, of the kind
that have been produced from Eocene and Cretac-
eous ‘glassy forams’ (Pearson et al. 2001; Wilson
& Norris 2001), therefore, are a priority for better
constraining global sea-surface temperatures and
E–O latitudinal thermal gradients.
Other temperature proxies – Much of the study
of E– O palaeoclimates is limited by our ability to
accurately determine temperature variability. As
H. K. COXALL & P. N. PEARSON360

Fig. 3. Eocene– Oligocene benthic foraminifera palaeoclimate proxy records from ODP Site 1218 examining possible
changes in sea level, ice volume and temperature (after Coxall et al. 2005). Data are plotted on the ODP Site 1218
orbitally tuned timescale. (a)d
18
O (crosses, 5-point moving average trend line) versus Mg/Ca (triangles, 3-point
moving average trend line), in principle an independent palaeothermometer (c. 35 to 31 Ma) from ODP Site 1218
(Mg/Ca data from Lear et al. 2004). The records show no decrease in Mg/Ca across the E –O transition and into the
EOGM. In fact, the Mg/Ca data show an increase from latest Eocene to earliest Oligocene suggesting either
bottom-water warming or that Mg partitioning into benthic foraminiferal calcite is sensitive to factors other than
temperature (e.g. increasing pH with CCD deepening, see Fig. 7). d
18
O temperatures shown apply to a world free of
continental-scale ice-sheets (d
18
O
w
¼21‰ Standard Mean Ocean Water (Kennett & Shackleton 1976). Equilibrium
calcite values ¼d
18
O
c
þ0.64‰) (Kennett & Shackleton 1976). (b) & (c) Estimated global ice budgets and
glacioeustatic sea-level fall associated with onset of Antarctic glaciation for ice with oxygen isotope values of 250‰
and 230‰, assuming that all of the d
18
O increase associated with Oi-1 is attributable to increased ice volume.
Arrows indicate modern Antarctic ice volume (c. 25.4 "10
6
km
3
) and apparent sea-level fall (ASL, defined as eustacy
plus the effects of water loading on the crust) (70 m) estimated for the Eocene –Oligocene Transition by sequence
stratigraphy (Pekar et al. 2002).
THE E–O TRANSITION 361

discussed above, the deep-sea climate proxy records
used routinely across the E– O interval that are temp-
erature sensitive are also affected by other factors
such as changing global ice volume and possibly
alkalinity related to CCD shift and or changes in
atmospheric pCO
2
. The range of additional indepen-
dent palaeotemperature proxies is limited for
this time interval because of preservation issues,
but there are several that provide information about
E–O changes especially in continental regions.
Palaeobotanical assemblage associations (see
above) and leaf margin analysis document signifi-
cant cooling in the Late Eocene to Early Oligocene
of Saxony, Germany (Roth-Nebelsick et al. 2004).
d
18
O variations in low latitude shallow-marine mol-
luscs and fish otoliths also indicate E-O cooling, and
increased seasonality (i.e. cooler winters) (Kobashi
et al. 2001; Ivany et al. 2000). Elsewhere in North-
ern Europe, summer palaeotemperature estimates
for continental freshwater derived from rodent
tooth enamel, molluscs and fish otoliths, suggest a
fluctuating mesothermal climate during Eocene
Oligocene time (Grimes et al. 2005) but no clear
evidence for climatic cooling. Cooling in this
record, however may be masked by the large
error bars associated with the mean fresh water
d
18
O estimates (5–6%.) because other continental
records derived from fossil bone and tooth enamel
indicate a large drop in mean annual temperatures
of c. 88C over 400 kyrs (Zanazzi et al. 2007). In
the latter study, continental cooling appears to
have been delayed in time with respect to the
marine changes by c. 400 kyrs. Sedimentary
records from the Tibetan Plateau provide additional
evidence for cooling and aridification in continental
Asia precisely at the time of the Eocene-Oligocene
Transition (Dupont-Nivet et al. 2007).
In the future, organic biomarkers, such asTEX
86
(TetraEther indeX of tetraethers with 86 carbon
atoms), a proxy for determining sea-surface temp-
erature from the fossilized membranes of marine
archaea (or Archaebacteria) (Schouten et al.
2002), may prove useful for obtaining accurate
temperature reconstructions. Recent recovery of
well-preserved organic biomarkers of early Palaeo-
gene age from Tanzania (van Dongen et al. 2006)
suggests that such temperature estimates are not
far off (Pearson et al. 2007).
Sea level – Widespread glaciation of Antarctica
at the E–O boundary, and possibly elsewhere,
would have resulted in a significant drop in global
sea level proportional to the magnitude of ice
growth. Identification and quantification of E–O
eustatic change, therefore, is important for under-
standing this climatic event. Although there are
doubts concerning the relative contributions of
temperature and ice volume to the d
18
O record, it
is generally considered that Antarctic ice volume
achieved at least 50% of its present development
at the E–O boundary (Barker et al. 1999;
DeConto & Pollard 2003a). Associated sea-level
fall has been estimated at 30 –90 m by Miller
et al. (1991) and modelled more conservatively at
40–50 m by DeConto & Pollard (2003a). The
classic Cenozoic sea-level curves, based on
sequence stratigraphy, failed to recognize sea-
level fall close to the E –O boundary and place
the largest change in the middle Oligocene (e.g.
Vail & Hardenbol 1979; Haq et al. 1987).
However, there are large uncertainties in dating
and correlation in these early reconstructions and
recent sequence stratigraphic studies on the New
Jersey continental margin have identified evidence
for a prominent eustatic lowering, indicating as
much as 70 m absolute sea-level fall, coincident
with the EOGM (Miller & Mountain 1996; Pekar
& Miller 1996; Pekar et al. 2002). Eustatic change
is further supported by covariance between
benthic and tropical planktonic d
18
O records
(Miller et al. 1991). Evidence of significant E–O
sea-level fall has been also been recorded on the
West African margin (Se
´ranne 1999), in the North
Sea region (Vandenberghe et al. 2003) and
Southern Australia (McGowran et al. 1992),
although these studies do not provide estimates of
the magnitude of change.
Surprisingly, E– O sea-level falldoes not appear to
be ubiquitous, and in other regions there is little or no
evidence for large-scale change, even in quasimarine
continental facies that should be very sensitive to
change. For example, estuarine sedimentary records
from the Hampshire Basin, UK, suggest maximum
relative sea-level fall of 15 m (Gale et al. 2006) and
in Mississippi, USA, the stratigraphy has been inter-
preted as indicating sea-level highstand at the E–O
boundary with evidence for increasing water depth
across the transition (Echols et al. 2003). These
studies question the extent of E– O sea-level change
and/or suggest that it coincided with widespread
tectonic activity that masked the full magnitude of
eustatic response. The lack of consensus and quanti-
tative agreement of benthic proxies is one of the
current areas of disagreement regarding the E–O
transition and there is still sckepticism among some
workers over whether there was any global change
in sea level, and thus, ice volume, at all (Hay et al.
2005).
Carbon cycling and productivity
Marine carbonate carbon stable isotopes –
Deep-sea benthic foraminifera and bulk carbonate
sediment d
13
C reflect regional and local changes
in surface to deep organic carbon cycling as well
as global changes in organic and inorganic carbon
H. K. COXALL & P. N. PEARSON362

burial as carbon moves between the lithosphere,
oceanic and atmospheric reservoirs. The commence-
ment of Oligocene glaciation was accompanied by
an oceanwide positive d
13
C anomaly of up to
1.0‰ (e.g. Zachos et al. 1996, 2001; Diester-Haass
& Zachos 2003; Coxall et al. 2005) (Fig. 4). The
shift into the d
13
C anomaly was also rapid (several
hundred kyr) and stepwise (Coxall et al. 2005),
with peak values recorded at the base of the
EOGM. ODP Site 1218 records show a distinctive
d
13
C ‘overshoot’ of typical early Oligocene
values followed by a longer recovery phase
(400–600 kyr) to near pre-excursion values
(Fig. 4b). The long recovery is probably a response
to the rapid deepening of the oceanic carbonate com-
pensation depth (see below) (Coxall et al. 2005;
Zachos & Kump 2005). In the high-resolution
record from ODP Site 1218, the d
18
O and d
13
C
step-shifts are similar in form, but d
13
C lags by
c. 10 kyr (Coxall et al. 2005). A lag between d
13
C
and d
18
O was also recognized in Site 522 records
(Zachos et al. 1996). This pattern had reversed by
the Oligocene because spectral analysis shows a
phase lag of d
18
O to d
13
C of c. 8 kyr with respect to
Fig. 4. Benthic foraminifer d
13
C from deep-sea drill sites across the Eocene –Oligocene Transition. (a) Compilation
(after Zachos et al. 2001) from mid- to high southern latitudes (diamonds) and the equatorial Pacific (other symbols)
plotted on a common timescale (Berggren et al. 1995) showing the globally recognizable c. 1‰ shift and positive
excursion. Note that the two figures are plotted on different timescales. The E–O boundary, as identified by Poore
(1984), is shown in relation to the Site 522 data (see Fig. 2). (b)d
13
C (crosses) from ODP Site 1218 (Coxall et al. 2005)
plotted on an orbitally tuned timescale and compared to d
18
O (5-point running mean from Fig. 2), the isotope ‘Shift’
and the EOGM. A stepped pattern of change, including an intermediate plateau, is clear in the Site 1218 d
13
C record.
There is an also an indication of step features in the compilation, especially Site 522.
THE E–O TRANSITION 363

the 40 kyr band, suggesting that the response of the
global carbon cycle may have helped force changes
in early Oligocene climate (Coxall et al. 2005).
One hypothesis regarding the carbon isotope
anomaly relates it to a temporary increase in the
ratio of organic carbon/CaCO
3
burial rates
because of enhanced marine export production,
brought about by climate-induced intensification
of wind stress, upwelling and oceanic turnover
(e.g. Shackleton & Kennett 1975; Diester-Haass
1992, 1995; Zachos et al. 1996 and references
therein – see below). In addition, the supply of
production-limiting nutrients, especially iron, is
likely to have increased in response to sea-level fall
and enhanced continental weathering, as has been
suggested during Quaternary glacial/interglacials
(e.g. Martin 1990). Observations are also compatible
with recent modelling studies that suggest the d
13
C
shift reflects reorganization of the carbon cycle in
response to a rapid drawdown of carbon dioxide
related to increased biological production and CCD
deepening (Zachos & Kump 2005). Qualitative and
quantitative proxy evidence for these parameters
and processes is presented in more detail below.
Marine productivity – Variations in palaeo-
marine productivity and export production provide
important information about global carbon cycling
and the causal/feedback roles they play in regulat-
ing climate. Significant effort has been focused on
obtaining proxy records of E–O productivity
changes. Commonly used methods include benthic
foraminifera and other marine micro- and macro-
fossil accumulation rates, opal/carbonate accumu-
lation, carbonate dissolution indices, carbonate
d
13
C and marine barite accumulation (e.g. Salamy
& Zachos 1999; Diester-Haass & Zachos 2003;
Averyt et al. 2005). Fossil accumulation rate
proxies are thought to vary in direct response to
standing biomass in the water column and phyto-
detritus availability at the sea floor but can be
biased by age model problems (Diester-Haass &
Zachos 2003). Carbonate dissolution indices
(based on foraminiferal shell fragmentation)
provide insight into primary production because
dissolution is affected by the rain ratio of organic
carbon/CaCO
3
(Diester-Haass & Zachos 2003).
Palaeoproductivity proxy data point towards a
sharp increase in export production associated
with the E–O transition, especially in the southern
high latitudes. Benthic and planktonic foraminifera,
radiolaria, fish debris, echinoderms and ostracodes
all show significant increases in accumulation rate
in parallel with the isotopic shifts into the EOGM,
in both Southern Ocean (Ehrmann & Mackensen
1992; Diester-Haass 1995, 1996; Diester-Haass &
Zahn 1996; Salamy & Zachos 1998) and mid- and
low-latitude regions (Diester-Haass & Zahn 2001;
Diester-Haass & Zachos 2003; Vanden Berg &
Jarrard 2004). Other reports, however, find no
evidence for increased productivity in equatorial
regions (Nilsen et al. 2003; Schumacher & Lazarus
2004) suggesting that there was considerable
regional variation in circulation and productivity
responses to the climate change outside of the
Southern Ocean region. The highest-resolution
records suggest that the increase in productivity
was initiated during the isotopic shift, i.e. during
the E–O transition, and was sustained for at least
the duration of the EOGM (Salamy & Zachos
1999; Diester-Haass & Zachos 2003).
Records from the Southern Ocean suggest that
net productivity increased by several-fold in
response to the EOGM (e.g. Baldauf & Barron
1990; Salamy & Zachos 1999) with evidence for
increased seasonality in the delivery of organic
matter to the sea floor (Thomas & Gooday 1996;
Schumacher & Lazarus 2004) that may have facili-
tated more rapid burial of organic carbon.
Calcareous primary producers, dominant in the
Late Eocene, were partially replaced by opaline
organisms suggesting a trend toward seasonally
greater surface divergence and upwelling. A corre-
sponding increase in organic carbon content has not
been widely recorded, but organic content is gener-
ally very low in Palaeogene-age deep-sea sediments
because of rapid post-burial consumption, and the
signal may not be preserved (Curry et al. 1995).
Increases in carbonate accumulation proxies is
complicated by the major deepening of the CCD
synchronous with the onset of the E –O shift, thus,
integration of multiple independent proxies is
important for differentiating productivity signals
from secondary processes, such as dissolution
related to lysocline shift.
Other productivity proxies such as barium, reac-
tive phosphorus and nitrogen isotopes may provide
complementary evidence for changes in export
production that are independent of carbonate dissol-
ution issues. For example, Nilsen et al. (2003) con-
cluded that although biogenic silica production/
preservation increased during the E–O transition
in equatorial Atlantic ODP Site 925, barium and
reactive phosphorus suggested no change in pro-
ductivity or nutrient burial, implying that the
increased opal production was relative and occurred
at the expense of carbonate production.
The cause of worldwide increased palaeoproduc-
tivity from the Late Eocene to the Oligocene is
attributed to ocean mixing/circulation changes that
increased the availability of nutrients in surface
waters. Whether this was driven by increased latitu-
dinal thermal gradients stimulating stronger winds,
upwelling and mixing, or reorganization of global
circulation related to the opening and closing of tec-
tonic ocean gateways is still a matter of much debate
(Diester-Haass 1996; Diester-Haass & Zahn 1996;
H. K. COXALL & P. N. PEARSON364

Salamy & Zachos 1999; Diester-Haass & Zahn
2001; Diester-Haass & Zachos 2003; Hay et al.
2005). Either way, the result of the widespread
palaeoproductivity increase, and a phase of
increased organic carbon burial, might have led to
a lowering of oceanic and atmospheric CO
2
that
could have played a key role in E– O climate
change by: (1) directly forcing the extreme transient
cooling and glaciation, and/or (2) providing a posi-
tive feedback that further contributed to global
cooling once the transition was initiated (Diester-
Haass & Zachos 2003; Zachos & Kump 2005).
Atmospheric carbon dioxide
Changes in the partial pressure of atmospheric
carbon dioxide (pCO
2
) is thought to be a primary
driver of climate over Phanerozoic time (Raymo
1991; Berner 1992; Crowley 2000) and especially
during transitions into and out of glaciations (e.g.
Barnola et al. 1987; Petit et al. 1999; Royer et al.
2004). General circulation models of Palaeogene
climate have been used to show that continuous
depletion of pCO2, amplified by Milankovitch
forcing and ice-albedo feedbacks, could cause
significant temperature reduction resulting in per-
manent continental ice-sheets in high latitudes
(e.g. DeConto & Pollard 2003a,b; Pollard &
DeConto 2003a; Zachos & Kump 2005). Recon-
structing E–O pCO
2
, therefore, is a primary goal
for constraining the links between climate and
radiative forcing of the Earth’s surface temperatures
through the E–O transition.
A variety of proxies are used to estimate pCO
2
during the early Cenozoic including; isotope ratios
of marine carbonates or soil carbonates; stomatal
density/index in fossil leaves; alkenone proxy in
organic biomarkers; boron isotopic ratios in fossil
surface-dwelling foraminifera (see Royer et al.
2001; Boucot & Gray 2001 for reviews). These are
discussed further below. In addition, models have
been used to simulate CO
2
variations over Phanero-
zoic time. Such models are based on changes in
carbon burial on land (due to the evolution of land
plants) and weathering changes (caused by organic
acids in soils) (e.g. Berner 1990) and suggest a
general decline in the early Cenozoic. Although
there is a good first-order agreement between
model results and proxy CO
2
estimates on long time-
scales (Berner 1990, 1997; Crowley 2000; Berner &
Kothavala 2001; Royer et al. 2007), the records
mostly lack the detail and accuracy to fully resolve
the magnitude and timing of decline during the
E–O transition and other key climatic events.
Fossil stomata proxies – The palaeobotanical
approach for determining past CO
2
concentration
is based on the inverse relationship between leaf
stomatal density and pCO
2
in some types of plants
(e.g. Beerling et al. 1993). Application of this
method using early Cenozoic plant fossils has
yielded conflicting results. Stomatal counts on the
gymnosperms Ginkgo biloba and Metasequoia
glyptostroboides suggest that pCO
2
was stable at
between 300 and 450 parts per million (ppm)
by volume during the Eocene and Neogene
(Royer 2001; Royer et al. 2001), whereas results
of a later study on Ginkgo suggest a decrease in
pCO
2
at the E–O boundary (Retallack 2002).
Stomatal densities in fossil leaves from three
dicotyledon angiosperm taxa (Eotrigonobalanus
furcinervis,Laurophyllum pseudoprinceps and
Laurophyllum acutimontanum) suggest that pCO
2
was higher during the Late Eocene than during the
Early Oligocene (Roth-Nebelsick et al. 2004).
These differing results indicate large uncertainties
in the method and probably different stomatal
responses to pCO
2
in different species. This is not
surprising because, although stomatal density may
decrease somewhat systematically in laboratory
experiments with increasing pCO
2
, the response
on evolutionary timescales is unknown and the cali-
bration curve used to calculate pCO
2
with the sto-
matal indices are subject to wide uncertainties
especially at higher than pre-industrial values
(Roth-Nebelsick et al. 2004). Consequently,
results using stomatal proxy methods should be
interpreted cautiously.
Boron isotopes – The boron-isotope ratios of
planktonic foraminifera have been used to estimate
the pH of surface-layer seawater, which is in turn
used to estimate pCO
2
(e.g. Pearson & Palmer
1999, 2000; Hs
ˇnisch & Hemming 2004). Results
suggest an erratic decline in pCO
2
from 3000–
4000 ppm between 55 and 40 Myr ago to
1000 ppm for the Late Eocene but are based on
various assumptions (see e.g. Pagani et al. 2005a).
Data for the E–O transition are lacking but values
are suggested to have decreased by at least a factor
of two by the Miocene (Pearson & Palmer 2000).
Alkenones –pCO
2
estimates using the carbon
stable isotope composition of marine alkenones
indicate a similar pattern of pCO
2
decline through
the Palaeogene and Neogene to the boron isotope
method, but there are also significant differences
(Pagani 2002; Pagani et al. 2005b). Results
suggest that pCO
2
ranged between 1000 to
1500 ppm in the middle to Late Eocene then
decreased in several steps during the Oligocene,
reaching ‘modern’ levels by the latest Oligocene.
In detail, the alkenone record shows that pCO
2
remained high across the E –O boundary and into
the lower Oligocene and began falling rapidly at
c. 32 Ma (Pagani et al. 2005b).
Palaeosol estimates –pCO
2
estimates based on
palaeosols isotopic data support a pattern of higher
THE E–O TRANSITION 365

concentrations during the Eocene and Oligocene
compared to the Neogene (e.g. Cerling 1991,
1999; Ekart et al. 1999; Lowenstein & Demicco
2006).
In summary, despite some discrepancies, inde-
pendent proxies largely support a pattern of higher
than modern (2 to 5 times) atmospheric pCO
2
in
the Late Eocene, falling through the Cenozoic and
reaching near-modern values in the lower
Miocene. Possible causes are reduced sea-floor
spreading (Berner 1990), increased silicate weath-
ering (Raymo & Ruddiman 1992; Ruddiman et al.
1997) and increased organic carbon burial (Diester-
Haass & Zachos 2003). Although there is a sugges-
tion of a trend towards decreasing pCO
2
through the
Eocene (Pearson & Palmer 2000; Pagani et al.
2005b), the alkenone data suggest that pCO
2
remained high (1000 –1500 ppm) until after the
E–O climatic transition, possibly following termin-
ation of the EOGM (Pagani et al. 2005b). This
questions the proposed relation between pCO
2
and
the initiation of glaciation on Antarctica (DeConto
& Pollard 2003a). However, none of the proxies
or age correlations are perfect (e.g. Pagani et al.
2005a; Roth-Nebelsick et al. 2005) and despite
the issues of timing it is still possible that the
pCO
2
decrease was a critical factor that allowed
expansion of ice-sheets on Antarctica.
Carbonate and the CCD
The carbonate compensation depth (CCD) is the
depth in the ocean at which the supply of carbonate
produced in surface waters is balanced by dissol-
ution. The CCD is sensitive to ocean fertility,
acidity and pCO
2
and, thus, is closely linked to
global climate. Cenozoic palaeodepth reconstruc-
tions showing the distribution of carbonate sedi-
ments across the ocean basins reveal a general
increase in deep-sea carbonate content in the Oligo-
cene compared to the Eocene associated with a
major global deepening of the CCD (Van Andel
1975; Thunell & Corliss 1986) (Fig. 5a). The dee-
pening was most extreme in the eastern equatorial
Pacific (estimated at 1200 m, Rea & Lyle 2005)
but has also been recorded in the Atlantic Ocean
(c. 1 km; Hsu
¨et al. 1984; Moore et al. 1984) and
the Indian Ocean (max. 700 m; Peterson &
Backman 1990), indicating that it was a global
phenomenon reflecting large changes in ocean
palaeochemistry. The effect of the CCD deepening
was to more than double the area of sea-floor
available for CaCO
3
deposition (Rea & Lyle 2005).
Recent CCD proxy records from equatorial
Pacific ODP Site 1218 have provided improved
constraints on the timing of the CCD shift and its
relation to the EOGM (Coxall et al. 2005). These
records show a very rapid increase in CaCO
3
content and CaCO
3
mass accumulation rate
(MAR) close to the E–O boundary (Fig. 5b). The
shift occurred in parallel with the d
18
O and d
13
C
increase, and shows the same two-step structure,
indicating that deepening CCD shift was synchro-
nous with the climatic shifts that preceded the
EOGM. Data and model simulations suggest that
under Pleistocene conditions, the increase in
deep-sea carbonate ion concentration associated
with a 1-km deepening of the global lysocline
yields a drawdown of atmospheric CO
2
of less
than 25 ppm (Sigman & Boyle 2000; Zeebe &
Westbroek 2003). Thus, E–O CCD deepening is
unlikely on its own to have caused glaciation
through the reduction of pCO
2
. More likely, the gla-
ciation triggered the CCD shift. The CCD probably
deepened to compensate for a reduction in the
global ratio of CaCO
3
to organic carbon burial, as
suggested by the calcite d
13
C increase. The
observed increase in the carbonate saturation state
of the deep ocean could have been achieved in
various ways, which may have acted in combi-
nation: (1) a shift in the locus of carbonate sedimen-
tation from the shelf to the deep sea (Berger &
Winterer 1974; Opdyke & Wilkinson 1989; Coxall
et al. 2005); (2) a rapid increase in the amount of
calcium entering the oceans (Rea & Lyle 2005);
(3) an increase in global siliceous (at the expense
of calcareous) plankton export production (e.g.
Thunell & Corliss 1986; Harrison 2000). The bio-
mineral shift hypothesis is supported by microfossil
evidence from the Southern Ocean and elsewhere
that demonstrates increased opal accumulation in
the early Oligocene (e.g. Diester-Haass & Zahn
1996; Salamy & Zachos 1999; Diester-Haass &
Zachos 2003). The relative importance of these
hypotheses can be tested using biogeochemical
models such as developed in Zachos & Kump
(2005).
Also of note is an interval of severe carbonate
dissolution in the top of magnetic Chron C13r
that pre-dates the CCD shift (Fig. 5b). At ODP
Site 1218, CaCO
3
content of the sediment and
CaCO
3
MAR fall to zero for an interval of
c. 0.6 m (#200 kyr) close to the base of the shift
into the EOGM. In fact, the precise timing of the
base of the d
18
O shift is lost in this record
because there are no benthic calcitic foraminifera
for analysis. A similar interval of dissolution, cor-
relative with upper Chron C13r, occurs at DSDP
Site 522 (Zachos et al. 1996), Equatorial Atlantic
ODP Site 925 (Diester-Haass & Zachos 2003)
and ODP Site 1265 (Liu et al. 2004; Tuo et al.
2006). This suggests a temporary shallowing of
the CCD immediately before the build-up to the
EOGM, the significance of which has yet to
be established.
H. K. COXALL & P. N. PEARSON366

Weathering and sediments
Changes in the flux, nature and chemistry of
deep-sea marine sediments indicate the style and
intensity of contemporaneous weathering and the
possible sources of incoming sedimentary material,
all of which vary with climate.
Glaciomarine sediments – The most direct evi-
dence of E– O Antarctic glaciation is the occurrence
of glaciomarine sediments, i.e. water-lain glacial tills
(including ice-rafted debris), sands and diamictites
(coarsely sorted boulder- to sand-sized grains), on
or close to Antarctica and elsewhere. These
sediments represent material that was scraped off
and entrained by moving ice and subsequently
deposited at the front of retreating glaciers or trans-
ported offshore by icebergs. The pattern of early
Cenozoic glaciomarine sedimentation as recovered
in deep-sea cores indicates several episodes of
expansive continental glaciation during the Oligo-
cene and several short-lived episodes in the Late
Eocene in east Antarctica (e.g. Barrett et al. 1989;
Wise et al. 1991; Breza & Wise 1992; Zachos et al.
1992, 1999). There is now also glaciomarine evi-
dence for a regionally extensive West Antarctica ice-
sheet of E– O age (Birkenmajer 1996; Gilbert et al.
2003; Birkenmajer et al. 2005; Ivany et al. 2006).
These records suggest that the Oligocene glaciation
was not restricted to East Antarctica and, thus,
imply an extreme climatic response to the forcing
Fig. 5. Deep-sea records showing changes in the Eocene– Oligocene equatorial Pacific Calcite Compensation Depth
(CCD). (a) Long-term Cenozoic record of CCD depth from 50 Ma to Present sediments based on carbonate mass
accumulation rate (MAR) (after Tripati et al. 2005). Black line is the CCD depth reconstructed from classic low
resolution Deep Sea Drilling Project records (modified from Van Andel 1975). Grey line is the revised E –O CCD
history based on ODP Leg 199 sediments. (b) High-resolution %CaCO
3
and CaCO
3
MAR (5-point running means)
across the E –O transition from ODP Site 1218 (palaeodepth c. 3800 m) (Coxall et al. 2005) on an orbitally tuned
timescale. The records show that CCD deepening (increase in CaCO
3
) occurred (i) faster than previously documented,
(ii) in two 40 kyr steps. The timing of CCD deepening is synchronous with the stepwise d
18
O shift into the EOGM
(Coxall et al. 2005).
THE E–O TRANSITION 367

factors that facilitated high-latitude ice expansion in
the earliest Oligocene (Ivany et al. 2006).
Information for the Northern Hemisphere is
limited. Results from recent drilling on Lomono-
sov Ridge, central Arctic Ocean, recovered the
first evidence of ice-rafted debris from the middle
Eocene, c. 35 Myr earlier than previously thought
(Moran et al. 2006). These sediments are inter-
preted as indicating Arctic cooling and significant
ice coverage coincident with Antarctic glaciation
and are used in support of arguments for bipolar
symmetry in early Cenozoic climate change (e.g.
Tripati et al. 2005). Understanding of the extent of
Arctic glaciation and pattern of climatic change,
however, remains limited because carbonate sedi-
ments, including calcitic micofossils that provide
important geochemical climate proxies, are missing
from Lomonosov cores, and the Late Eocene and
Oligocene fall in a hiatus (Moran et al. 2006).
Further drilling, therefore, is necessary to corrobo-
rate the initial findings from Lomonosov Ridge and
obtain appropriate stratigraphic coverage of the
E–O Transition. Recent reports of stratigraphically
extensive ice rafted material in the Late Eocene to
early Oligocene from the Norwegian-Greenland
Sea add support to the idea of contemporaneous
Northern Hemisphere glaciation (Eldrett et al.
2007). However, these records cannot determine
the extent of glaciation, i.e. whether the material
was deposited by icebergs calving off small ephem-
eral glaciers or larger ice-sheets.
Strontium isotope ratios – Variations in marine
carbonate
87
Sr/
86
Sr during the Cenozoic represent
changes in the balance of Sr inputs from hydrother-
mal flux and continental weathering (e.g. Palmer &
Edmond 1992; Richter et al. 1992; Reilly et al.
2002). Therefore, as well as providing a useful stra-
tigraphic tool (e.g. Burke et al. 1982),
87
Sr/
86
Sr has
the potential to record information on various pro-
cesses relevant to climate change, especially con-
cerning weathering shifts related to longer-term
variations in atmospheric CO
2
. The pattern of post-
Eocene
87
Sr/
86
Sr shows a general increase in values
with a series of stepped increases in the rate of Sr
isotopic variation superimposed. The first of these
steps occurs around the Late Eocene to early Oligo-
cene (Hess et al. 1986; Miller et al. 1991; Mead &
Hodell 1995; Zachos et al. 1999; Reilly et al. 2002).
Although changes in hydrothermal flux and carbon-
ate weathering are thought to have contributed to
the general
87
Sr/
86
Sr increase over the last 100
myr, increased riverine inputs due to enhanced con-
tinental weathering are believed to have been
responsible for the larger-scale isotopic variation
(e.g. Reilly et al. 2002; Lear et al. 2003).
The cause of the inferred enhanced weathering
is controversial. This is in part because of
differences in identifying the timing of the Sr
shift (see Reilly et al. 2002) and has been attribu-
ted to: tectonic uplift of the Himalayan Plateau,
which increased elevation and altered exposure of
rock lithologies (e.g. Raymo & Ruddiman 1992);
initiation of Antarctic glaciation, whereby
increased ice-sheet growth resulted in increased
mechanical and chemical weathering of continental
rocks (Miller et al. 1991; Zachos et al. 1999;
Zachos & Kump 2005); or a combination of
these processes (Mead & Hodell 1995; Reilly
et al. 2002).
Clay mineralogy – Variations in clay mineral-
ogy indicate changes in depositional environment
and provide insight into erosional processes
related to climate. Cenozoic clay mineral records
show a general trend towards illite –chlorite-domi-
nated assemblages and a reduction in smectite and
kaolinite. This is thought to reflect an increase in
the rock-derived supply and a decrease in soil-
derived supply of clay minerals to the ocean
(Chamley 1986), although the pattern is also
affected by local conditions. The shift begins in
the Late Eocene and has been recorded in marine
sediments globally (Chamley 1986; Ehrman &
Mackensen 1992; Diester-Haass et al. 1993;
Robert et al. 2002; Gale et al. 2006). Close to
Antarctica, the illite–chlorite shift has been inter-
preted as recording a switch from a climate typified
by alternating wet and dry seasons resulting in
predominantly chemical erosion, to intensified
physical (mechanical) weathering by glaciers on
eastern Antarctica under a cooler climate
(Ehrman & Mackensen 1992; Diester-Haass et al.
1993; Zachos et al. 1999; Robert et al. 2002). Ant-
arctic environmental magnetic records support this
hypothesis. For example, Ross Sea sediments show
a decrease in detrital magnetite during the early
Oligocene, which is interpreted as indicating a
shift to a cooler, drier climate leading to reduced
chemical weathering of igneous rocks on
Antarctica (Sagnotti et al. 1998; Wilson et al.
1998). The environmental magnetic records and
clay mineral behaviour is thus consistent with
other evidence for widespread cooling and glacia-
tion of Antarctica in the earliest Oligocene.
These records also constrain the timing of glacial
activity on Antarctica in the Late Eocene and
suggest intensification of glaciation in the basal
Oligocene.
Osmium isotopes – Os isotope variation in
seawater (
187
Os/
188
Os), as recorded by marine
sediments, is believed to reflect the contribution
of Os to the oceans from continents (rivers) and sub-
marine alteration, and the influx of extraterrestrial
material (cosmic dust). The long-term Cenozoic
record of
187
Os/
188
Os variation shows a pattern of
H. K. COXALL & P. N. PEARSON368

increasing values that was probably caused by some
combination of changes in weathering or increased
influx of cosmic dust (Peucker-Erenbrink et al.
1995; Pegram & Turekian 1999; Peucker-Erenbrink
& Ravizza 2000). The increasing numbers of
detailed Os isotope records across the E –O
transition provide good evidence for the global
influence of glaciation on the supply of Os to the
ocean (Ravizza & Peucker-Erenbrink 2003; Dalai
et al. 2006).
Bulk sediment
187
Os/
188
Os records from the
Atlantic and Pacific reveal a c. 30% stepwise shift
towards higher values in the early Oligocene after
a minimum in the Late Eocene (Pegram & Turekian
1999; Ravizza & Peucker-Erenbrink 2003; Dalai
et al. 2006). The stepwise shift is interpreted as
coinciding with the growth and decay of major ice-
sheets. Comparison of E–O
187
Os/
188
Os records
with high-resolution benthic foraminiferal d
18
O
from Site 1218 (Coxall et al. 2005) suggests that
Os flux to the oceans decreased during cooling
and ice growth, whereas subsequent decay of
ice-sheets and deglacial weathering drove seawater
187
Os/
188
Os to higher values, i.e. higher
187
Os/
188
Os values occur following termination of
the EOGM. The post-EOGM
187
Os/
188
Os shift is
interpreted to represent an increase in radiogenic
Os to the oceans derived from the weathering of
easily erodible glacial moraines following termin-
ation of the glaciation (Ravizza & Peucker-
Ehrenbrink 2003; Dalai et al. 2006). It is suggested
that
187
Os/
188
Os remained high through the Oligo-
cene because subsequent cyclic variations in ice
volume regularly replenished sediment supply
(Ravizza & Peucker-Ehrenbrink 2003).
Dust – Aeolian grains in pelagic clays, i.e.
wind-blown dust particles, show an increase in
mean grain size beginning in the Late Eocene
close to the E–O boundary. This has been attributed
to more vigorous atmospheric circulation associated
with the climatic reorganization (Rea et al. 1985;
Ravizza Peucker-Ehrenbrink 2003; Vanden Berg
& Jarrad 2004). A more controversial idea is that
cosmic dust flux in the Late Eocene may have con-
tributed to global cooling and the EOGM by supply-
ing bio-essential trace elements to the oceans and
thereby resulting in higher ocean productivity,
enhanced burial of organic carbon and drawdown
of atmospheric CO
2
(Dalai et al. 2006).
The sedimentary and geochemical evidence for
changes in weathering associated with the E –O
transition is consistent with other proxies that indi-
cate widespread glaciation. Weathering-CO
2
climate feedbacks undoubtedly play a role in
forcing or moderating climate but their importance
is widely debated (Volk 1987; Kump et al. 2000;
Lear et al. 2004).
Late Eocene impacts
The discovery of three bolide impact craters
(Chesapeake Bay, Toms Canyon and Popigai; e.g.
Poag 1995; Bottomley et al. 1997) and associated
breccia and ejecta deposits in the Late Eocene has
provoked speculation that the biotic turnover and
climatic cooling associated with the E –O boundary
was triggered by the impact of an extraterrestrial
body or bodies (e.g. Keller 1986; Vonhof et al.
2000). Recent age correlation of the impact sites
shows that they all occurred within magnetic
Chron C16n.2n (Poag et al. 2003), which precedes
the E–O boundary by about 2 million years, and
may have formed part of a ‘comet shower’ (Poag
et al. 2003). Minor extinction episodes and assem-
blage shifts among planktonic foraminifera (e.g.
Keller 1986) and dinocysts (Vonhof et al. 2000)
and a negative d
13
C excursion associated with the
impact horizons at several sites (Poag et al. 2003)
has been taken as evidence for global-scale long-
term enviornmental disturbance related to the
impacts. However, it seems unlikely that the
impacts had a major influence on the Earth’s
biosphere at the E–O boundary and they are
discounted here as a possible cause of the climatic
transition. It is possible, however, that climatic
disruption caused by atmospheric dust loading
associated with the Late Eocene impacts, coupled
with an ice-albedo feedback mechanism that ampli-
fied impact-induced climatic cooling, may have
contributed to global cooling in the run-up to the
E–O transition (Vonhof et al. 2000).
Cyclostratigraphy and Milankovitch
cyclicity
Orbital dynamics that affect the amount and distri-
bution of incoming solar radiation received by the
Earth are thought to have been the principal ‘pace-
maker’ of Quaternary climate cycles, forcing the
repeated growth and decay of continental glaciers
(Imbrie et al. 1984). The effects of orbital fluctu-
ations in the early Cenozoic, however, when
climate boundary conditions differed and global
ice volumes may have been significantly less than
in the Quaternary, have been difficult to resolve.
One problem has been uncertainties in tracing astro-
nomical solutions that predict the position of indi-
vidual cycles back into the Palaeogene (Laskar
1999). However, much progress has been made
recently, and astronomical models have been
extended back to the early Oligocene and beyond
(Laskar et al. 2004; Pa
¨like et al. 2004), allowing
orbital tuning of E–O records (Coxall et al. 2005;
Gale et al. 2006; Jovane et al. 2006).
THE E–O TRANSITION 369

Improved deep-sea core recovery and automated
quasi-continuous core and downhole logging analy-
sis have resulted in a number of E–O climate proxy
records, including magnetic susceptibility, percent
carbonate and benthic d
18
O and d
13
C, of suff-
icient resolution for resolving patterns of cyclic
variation on Milankovitch timescales (e.g. Mead
et al. 1986; Hartl et al. 1995; Zachos et al. 1996;
Coxall et al. 2005; Jovane et al. 2006; Gale et al.
2006). Changes in surface productivity and bottom-
water redox conditions in response to climate are
thought to be responsible for the observed periodi-
city in sediment physical property records,
whereas fluctuating ice volumes and behaviour of
the global carbon cycle are thought to be the
major climatic control on d
18
O and d
13
C respect-
ively (Zachos et al. 1996).
Spectral analysis of ODP Site 1218 d
18
O and
d
13
C records, which are the highest-resolution E– O
time series available, show that all Milankovitch
periods are encoded (i.e. c. 20, c.40, c.100 and
c. 400). Power is concentrated at the obliquity
(40 kyr) frequency for d
18
O and at the 400-kyr
eccentricity frequency, with a weaker 40 kyr com-
ponent for d
13
C (Coxall et al. 2005). A similar
c. 40 kyr spectral peak was recognized in
DSDP Site 522 d
18
O records (Zachos et al. 1996).
In the Oligocene, a phase lag of d
18
O with respect
to d
13
C of c. 8 kyr in the 40 kyr band suggests
that the response of the global carbon cycle
helped to force climate changes in early Oligocene
climate rather than reacting to it (Coxall et al.
2005). As in Neogene records, the clear c. 40 kyr
spectral peak in d
18
O is probably associated with
changes in Earth’s tilt axis and is consistent with
a strong high-latitude (ice volume) influence on
global climate (e.g. Imbrie et al. 1984). A new
13-million-years-long record from ODP Site 1218
for the entire Oligocene, which includes the E–O
dataset, confirms this pattern and identifies
405, 127 and 96 kyr Earth’s eccentricity and
1.2-million-year obliquity cycles pacing period-
ically re-occurring glacial and carbon cycle events
throughout the Oligocene (Pa
¨like et al. 2006).
Time-series analysis of other datasets shows
similar patterns. Spectral analysis of astronomically
tuned terrestrial illitic clay data from the Hampshire
Basin, UK, identifies strong c. 400 kyr eccentricity
and 41 kyr obliquity frequencies, with a smaller
peak at around 100 ka (Gale et al. 2006). Illite is
interpreted to have formed in palaeosols through
repeated wetting and drying in response to high sea-
sonality. The obliquity frequency cyclic variation
probably reflects changes in sea level driven by
minor fluctuations in the volume of Antarctic ice.
Longer-term environmental magnetic records from
the E–O stratotype in Massignano, Italy, reveal
cycles of c. 405 kyr (Jovane et al. 2004, 2006) as
a robust feature through the Late Eocene and into
the Oligocene. External forcing by orbital par-
ameters at eccentricity frequencies has also been
detected in E–O Southern Ocean (Diester-Haass
& Zahn 1996) and low-latitude productivity
proxies (Nilsen et al. 2003), suggesting that
orbital forcing caused a sedimentary response to
global climate change from at least Late Eocene.
The dominance of the long Milankovitch
periods in these records is unexpected because the
climate system response predicted by insolation cal-
culations based on the traditional Milankovitch
summer insolation hypothesis (Loutre et al. 2004)
is for the dominance of climatic precession and
obliquity, with only a very minor contribution by
Earth’s eccentricity periods. Modelling experiments
using a carbon cycle box model coupled to models
with orbital variations suggest that the explanation
for this phenomenon at least for d
13
C records is
amplification of the 405 kyr cycle due to the long
residence time of carbon in the oceans (c. 0.1
million years) (Pa
¨like et al. 2006).
In summary, time-series analyses performed on
early Oligocene datasets have revealed that 40 and
400 kyr Milankovitch cycles are a pervasive
feature of isotopic and other climate proxy
records, as in much of the Neogene (e.g. Zachos
et al. 1997; Paul et al. 2000), and further support
the presence of ice-sheets responding to orbital fluc-
tuations during this time period.
Deep-water circulation
The history of deep-water formation and circulation
provides important information about global
thermal gradients and circulation regimes. Today,
and for much of the later Cenozoic, the near-
freezing conditions at high northerly and southerly
latitudes produce the most dense in situ water
masses that sink to fill the deep ocean basins.
During the early Cenozoic, however, when the
high latitudes were warmer, the Southern Ocean
and even mid-latitudes are thought to have played
greater roles in deep-water formation (e.g. Miller
& Fairbanks 1985; Kennett & Stott 1990; Wright
& Miller 1993). Considerable effort has been put
into trying to reconstruct E –O deep-water circula-
tion history using a variety of different techniques.
Early reconstructions that utilized d
18
O (tempera-
ture), d
13
C (water mass age) and sediment distri-
bution suggest that the Southern Ocean was the
dominant source of deep water for the Late
Eocene, with pulses of bipolar deep-water
formation, i.e. an additional Northern Hemisphere-
sourced deep-water component, from the earliest
Oligocene (e.g. Wright & Miller 1993 and refer-
ences therein). This is supported by sedimentary
H. K. COXALL & P. N. PEARSON370

and seismic evidence that indicates the occurrence
of early Oligocene-age sediment drift deposits
reflecting a southern-flowing deep-water mass in
the North Atlantic (Kidd & Hill 1987; Davies
et al. 2001) and North Pacific (Scholl et al. 2003).
Deep-sea benthic foraminifera also provide indi-
cations of bottom-water conditions. The gradual
faunal changes in lower bathyal benthic for-
aminiferal assemblages observed in the Southern
Ocean and elsewhere have been interpreted as
suggesting that the cold, high-latitude-sourced
deep waters became established gradually as a
result of progressive cooling of surface waters at
high latitudes (Thomas 1992; Takata et al. 2007).
Neodymium isotopes – Additional evidence for
northern-sourced deep water is provided by neo-
dymium isotopes. Nd isotopic variations in sea-
water (
1
Nd
) reflect local Nd supply to the ocean
through weathering of continental rocks. Deep-
water masses sinking in a particular region therefore
carry a
1
Nd
signature characteristic of the proximal
Nd source (e.g. Goldstein & Jacobsen 1988; Via &
Thomas 2006 and references therein). This signa-
ture is preserved in the teeth and bones of fossil
fish (e.g. Wright et al. 1984) and can be extracted
to provide a useful tracer of deep-water influences
on geological timescales. Nd isotope data from
E–O sediments indicate that the initial transition
to a bipolar mode of deep-water circulation, invol-
ving the onset of Northern Component deep-water
formation, occurred in the early Oligocene (Scher
& Martin 2004; Via & Thomas 2006). Such data
seem to imply substantial cooling and/or ice
growth in the Northern Hemisphere at the climax
of the E–O transition.
Mechanisms and modelling
The cause of the E–O climate transition is still
widely debated and will doubtless remain so,
because such complex shifts in the Earth System
probably do not have a single identifiable cause.
The most widely held explanations for the funda-
mental climatic shift and initiation of the EOGM
are that it was caused by (i) cooling of Antarctica
due to changes in ocean circulation controlling
poleward heat transport, (ii) a gradually forced
threshold response to declining atmospheric
carbon dioxide levels. New drill sites and high-
fidelity proxy records are important for exploring
these hypotheses by documenting patterns of
change across the planet. Computer models,
however, which are becoming ever more sophisti-
cated in their ability to couple ocean, atmosphere,
cryosphere and biosphere climatic parameters,
have a valuable role in hypothesis testing and
allow exploration of climate behaviour that includes
a range of complex forcing and feedback mechan-
isms operating on global scales. Here the high-
quality proxy records provide essential boundary
conditions for constraining the models. The follow-
ing section summarizes evidence for and against the
possible mechanisms of E–O climate change
including the results of modelling studies, but we
emphasize that there is still more that we ‘don’t
know’ than that which we ‘know’.
Ocean gateways – One classical hypothesis
suggests that E–O glaciation was achieved
through thermal isolation of Antarctica following
the tectonic opening of ocean gateways between
Antarctica and Australia (Tasmanian Passage) and
Antarctica and South America (Drake Passage),
which led to the organization of the Antarctic
Circumpolar Current (ACC) (Kennett et al. 1975;
Kennett 1977; Murphy & Kennett 1986). This
idea draws on foraminiferal isotopic evidence
from deep-sea drill cores, principally from the
Tasman Sea, that suggests a shift from warm to
cold currents close to the E–O boundary and a
shift from calcareous to biosiliceous microfossils
in the early Oligocene. The theory has been
extended recently following a new campaign of
deep-sea drilling in the Tasman Sea region (Ship-
board Scientific Party 2001; Exon et al. 2004).
Renewed support for the hypothesis that thermal
isolation of Antarctica was the principal cause of
E–O climate change is also provided by the
results of (GCM) simulations that suggest the
opening of Drake Passage and the organization of
an ACC would have reduced southward oceanic
heat transport and facilitated cooling of Southern
Ocean sea-surface temperatures (Toggweiler &
Bjornsson 2000; Nong et al. 2000).
This view, however, remains controversial. An
alternative interpretation of siliceous microfossil
distribution from the region, principally dinocysts
and diatoms that are abundant in productive polar
regions, suggests that the Tasman Gateway
opened two million years before permanent glacia-
tion at about 35.5 Ma and that, although possibly a
factor contributing to the changes, this tectonic
event probably was not the cause of the E –O
climatic shift (Stickley et al. 2004). Moreover, an
alternate coupled ocean–atmosphere GCM chal-
lenges the idea that the opening of the Tasmanian
Passage contributed to Antarctic thermal isolation
by interrupting the southward flow of warm tropical
waters, because model simulations suggest that
warm waters did not penetrate to high latitudes in
the first place (Huber et al. 2004). This interpret-
ation is supported by another set of GCM simu-
lations which suggest that the opening of Southern
Ocean gateways plays a secondary role in this tran-
sition, relative to pCO
2
concentration (DeConto &
Pollard 2003a,b). Similarly, in an assessment of
THE E–O TRANSITION 371

the potential influence of the ACC on Antarctic
glaciation, it was concluded that the presence of
biosiliceous biofacies in the early Oligocene
cannot be used as evidence for the existence of a
continuous deep-reaching front, in the form of an
ACC, and that colder SSTs were an effect of glacia-
tion rather than the cause (Barker & Thomas 2004).
According to this study, Eocene–Oligocene dee-
pening of the Tasmanian gateway and timing of
glaciation are coincidental.
Precise timing of the opening of these tectonic
gateways is necessary for assessing their import-
ance as possible mechanisms of E–O climate
change. However, there are differences and uncer-
tainties in the age estimates of events, especially
for the opening of Drake Passage. Using tectonic
plate reconstructions, Lawver & Gahagan (2003)
suggested that a deep-water passage opened in the
Tasman Seaway close to the E–O boundary
(c. 33 Ma), and that the Drake Passage may have
been open by 31 Ma. The authors recognize a
major change in circulation at this time, with
increased production of Antarctic Bottom Water,
but they are cautious about linking this change
directly to the early Oligocene glaciation because
there is not a great deal of evidence to support
vigorous ACC until the Miocene (Rack 1993).
Using geophysical-modelling techniques,
Livermore et al. (2005) and Eagles et al. (2006)
estimate Drake Passage opened near to the E–O
boundary (34–30 Ma) and suggest that this tectonic
event was the trigger for rapid onset of early Oligo-
cene glaciation. Geochemical proxy data (Latimer
& Fillipelli 2002) and opal accumulation records
(Diekmann et al. 2004) from the Agulhas Basin
have also been interpreted as evidence for Drake
Passage opening to deep ACC flow close to
the E –O boundary at around 32 –33 Ma. The
Livermore et al. (2005) kinematic model predicts
that initial opening of the Drake Passage occurred
in the middle Eocene, which is considerably
earlier than previous estimates, and suggest that
flow of deep water through a series of restricted
basins may have contributed to gradual cooling
through the middle Eocene.
Nd isotopes also provide constraints on the
timing of Drake Passage opening. Scher and
Martin (2006) interpret Nd records from Agulhas
Ridge as indicating an influx of shallow Pacific
water in the late middle Eocene at c. 41 Ma,
suggesting that a connection through Drake
Passage was already open and probably facilitating
complete circum-Antarctic circulation by the Late
Eocene, i.e. before the opening of the Tasman
Seaway. Based on this timing, the authors conclude
that Drake Passage opening could not have initiated
the EOGM glaciation and propose circulation/pro-
ductivity linkages as a mechanism for declining
atmospheric carbon dioxide that may have been
the primary cause.
Despite recent and ongoing research into these
questions, opinions on the role of circulation
changes in the initiation of E –O climate change
remain divided. There appears to be a variety of evi-
dence to suggest that Southern Ocean gateways had
opened by the E–O boundary and that some form of
proto-ACC, which might have had a direct impact
on Antarctic climate, had been established by the
earliest Oligocene. Alternatively, new interpret-
ations of microfossil records and alternative GCM
simulations suggest that circulation changes
leading to thermal isolation of Antarctica did not
play a primary role in initiating glaciation. Consid-
ering these issues, Barker and Thomas (2004)
outline ideas for future work. They suggest that
future efforts should be focused on obtaining: (i)
better definition of the properties of oceanic
fronts, (ii) Cenozoic atmospheric pCO
2
proxy
records, (iii) further developments in models of
Antarctic glaciation, (iv) marine geology studies,
including grain-size, geochemical and mineralogi-
cal as a way of examining bottom-water circulation
patterns, and (vi) reappraisal of palaeontological
records previously interpreted as evidence for an
ACC.
Declining carbon dioxide and orbital
geometry – The second widely held hypothesis is
that E–O Antarctic glaciation was triggered by a
threshold response to declining atmospheric pCO
2
,
consistent with proxy records (see above), in com-
bination with a planetary orbital configuration that
favoured accumulation of annual snow. Modelling
experiments have bolstered support for this hypoth-
esis. In one set of simulations, sudden glaciation
was forced by gradually declining pCO
2
(DeConto
& Pollard 2003a). The results suggest that a
threshold may be reached whereby ice-sheet
height/mass-balance feedbacks cause isolated
ice-caps to expand rapidly, eventually coalescing
into a continental-scale ice-sheet (Fig. 6). Compari-
son of ODP Site 1218 proxy records with an orbital
template reveals that the beginning of the shift into
the EOGM coincides with a rare orbital configur-
ation comprising a phase of low eccentricity and
low-amplitude change in obliquity, favouring cool
austral summers and permitting accumulation of
summer snow that probably tipped the balance
towards glaciation (Coxall et al. 2005) (Fig. 7).
Further model sensitivity experiments suggest that
the range of atmospheric CO
2
variability needed
to induce such a rapid transition in the presence of
orbital forcing is 2 to 4 times pre-industrial levels
(Pollard & DeConto 2005). The pattern of step-
change in geochemical proxy records substantiates
the model prediction of a non-steady mode of
ice-sheet growth but the magnitude of the d
18
O
H. K. COXALL & P. N. PEARSON372

response is underestimated by the model (Coxall
et al. 2005).
A biogeochemical modelling study that was
used to investigate the role of the carbon cycle in
the glaciation process also supports a drawdown
of CO
2
as a likely initial forcing mechanism
(Zachos & Kump 2005). The models were able to
reproduce the observed positive excursions in the
mean d
13
C of inorganic marine carbon and biogenic
sediment accumulation rates by simple CO
2
balan-
cing feedbacks in the climate system associated
with ice-sheet coverage, silicate weathering rates,
and increased global carbon burial. However, the
post-EOGM timing of the stepped reduction in
pCO
2
, as suggested by recent proxy records,
questions this proposed relation between pCO
2
and the initiation of Antarctic glaciation (Pagani
et al. 2005b).
Additional forcing and feedback mechanisms –
A range of additional climate models is providing
further insight into the importance of other forcing
and feedback mechanisms on E–O climate
Fig. 6. A General Circulation Model simulation of the glacial inception and early growth of the East Antarctic Ice
Sheet (reprinted from DeConto & Pollard (2003a) with permission from authors and Nature Publishing Group).
(a) Model simulation of the transient climate-cryosphere response to a prescribed decline in CO
2
from 4 times to
2 times the pre-industrial atmospheric level over a 10-million year period. Ice volume, equivalent changes in sea
level (Dsea level) and the mean isotopic composition of the ocean (Dd
18
O) are shown for two pairs of simulations: (i)
Drake Passage (DP) open; (ii) DP closed (global climate model ocean heat transport coefficient increased in the
Southern Hemisphere). Equivalent sea-level changes were calculated according to the global ocean-area fraction
in the 34-million-year palaeogeography (0.731). Dd
18
O was calculated assuming a 0.0091 change in d
18
O per 1 m
change of sea level. (b) Modelled ice-surface elevations for four time slices during the E–O transition in a
10-million-year simulation (DeConto & Pollard 2003a). The simulation predicts that as CO
2
declined through the
Palaeogene, the gradual lowering of the Antarctic snowline began intersecting areas of high topography, first
producing small, isolated ice-caps (4.7 and 5.2 million years). As CO
2
continued to fall, height/mass-balance
feedbacks were initiated suddenly, producing larger dynamic ice-sheets that alternately coalesced and separated in
response to orbital forcing (5.8 million years). With further decline of CO
2
, a single, large EAIS became a
more permanent feature, almost insensitive to orbital forcing, with very little summer melting, and accumulation
zones reaching sea level around most of the continent (6.0 million years).
THE E–O TRANSITION 373

change. For example, modelling has suggested that
changes in Antarctic vegetation (i.e. needle leaf
forest to tundra) could have caused positive feed-
backs through changes in albedo that would have
played a significant role in the rapid glaciation
(Thorn & DeConto 2006). Other experiments have
shown that geothermal heat flux could strongly
influence aspects of Cenozoic Antarctic evolution
such as ice-sheet basal hydrology, sediment defor-
mation and discharge that appear to increase
orbital variability of ice-sheet behaviour (DeConto
& Pollard 2003a,b; Pollard et al. 2005). Models
have also suggested that non-linear Antarctic
ice-sheet transitions involving hysteresis have
played important roles in many of the observed pat-
terns of d
18
O change (Pollard & DeConto 2005).
The hysteresis effect in this context relates to the
sudden transitions in ice-sheet size between mul-
tiple stable states as a geometric consequence of
the intersection of ice-sheet surfaces with typical
Fig. 7. ODP Site 1218 glacial and CCD proxy records during the Late Eocene to Early Oligocene showing the
orbital pacing of Eocene–Oligocene climate change. Comparison of the records indicates that stepwise growth
of Antarctic ice-sheets (benthic foraminifera d
18
O increase), was synchronous with CCD deepening in the
equatorial Pacific and occurred during an eccentricity minimum and interval of low-amplitude obliquity change (grey
shading) favouring cool summers (Coxall et al. 2005).
H. K. COXALL & P. N. PEARSON374

spatial patterns, such as topography and coastline
geography (Pollard et al. 2005). The results of
these simulations suggest that the E–O global
climate shift may have been sensitive to the particu-
lar Antarctic conditions operating at the time. Even
so, it would be wise to consider Northern Hemi-
sphere changes as well (e.g. Moran et al. 2006;
Eldrett et al. 2007; DeConto & Pollard 2007).
Change in ocean structure and circulation – An
alternate view of E–O climate change is presented
by Hay et al. (2005). These authors suggest that the
transition from an Earth without perennial polar ice
to one with quasi-permanent polar ice could have
taken place rapidly only if both polar regions
became extensively ice-covered in a short time.
Since there is currently only limited evidence for
rapid E–O Northern Hemisphere glaciation, Hay
et al. (2005) suggest the global transition to an ice-
dominated climate system must have been gradual.
Their explanation relates the observed changes in
climate proxy data to changes in ocean structure
and deep-water circulation, specifically attributing
the E–O d
18
O shift to three factors: (1) a major
component of the d
18
O shift is related to tempera-
ture, representing the filling of the deep ocean
with cold water and development of a ‘psychro-
sphere’ (Kennett & Shackleton 1976); (2) a second-
ary part is attributed to the formation of permanent
sea-ice in the Arctic and perhaps around Antarctica,
with only minor growth of an East Antarctic
ice-sheet; (3) a large part of the d
18
O shift change
may be related to changes in alkalinity and decrease
in the pH of seawater as a result of uptake of CO
2
by
the ocean from the atmosphere. This hypothesis
rests largely on the lack of contemporaneous E –O
sea-level change interpreted from earlier sea-level
reconstructions but loses power in light of recon-
structions that link sequence boundaries implying
sea-level fall with global d
18
O increases (e.g.
Pekar et al. 2002).
Synthesis
The weight of fossil and climate proxy evidence
strongly indicates the Late Eocene to Early Oligo-
cene as a time of major global climate change,
with many implications for marine and terrestrial
ecosystems. Unlike the Cretaceous–Palaeogene or
Permian–Triassic boundaries, there is no indication
of a sudden catastrophic event; rather, fossil records
document a pattern of enhanced yet gradual turn-
over signalling adjustment to changes in food and
nutrient availability, habitat and climate regime.
Terrestrial biotas suggest significant and varied
environmental change across the latitudes, including
cooling, increased aridity and increased seasonality.
These changes are consistent with a scenario of
reduced thermal gradients, strengthening of zonal
winds and changes in precipitation patterns that
might result from cooling of one or both poles.
Most floral, and other terrestrial, fossil records,
however, lack the time resolution to correlate them
precisely with the EOGM as recorded by marine
records. Evolutionary turnover is recorded in many
terrestrial organisms, with highly varied ecologies,
which indicates a variety of external-forcing mechan-
isms. The clear signal from the terrestrial realm is that
there was no ‘Terminal Eocene Event’. Instead, the
transition was marked by a series of gradual extinc-
tions and speciations, before and after the E– O
boundary, that suggest interplay of environmental
selection pressures at variable intensities.
Patterns of evolutionary turnover in the pelagic
realm suggest changes to ocean thermal structure,
food availability and ocean temperatures. Phyto-
plankton show patterns of gradual turnover and
replacement by cosmopolitan, cold-tolerant
species. Tropical micro-zooplankton experienced
more extinction than other groups, although not
on a wide scale, and there was significant reduction
in tropical plankton diversity, especially among the
radiolaria. These records support the hypothesis of
increased oceanic turnover, a reduction in tropical
pelagic niches and more intense high-latitude,
equatorial and coastal upwelling associated with
the climatic shift into the EOGM. These records
can be tightly correlated with the proxy records of
Antarctic glaciation and show extinctions within
the isotopic shift that precedes the EOGM (i.e.
planktonic foraminifera and shallow-water
benthics), as well as major changes in plankton
communities during the EOGM. Deep-sea benthic
communities appear to have survived relatively
unchanged. Some even benefited from increased
food availability because of increased ocean
mixing and higher surface ocean productivity.
Diversification of large plankton-feeding whales,
in parallel with increased opal accumulation,
suggests that other elements of the pelagic ecosys-
tem were responding to an expanding nutritional
opportunity that arose during the EOGM.
Shallow marine environments appear to have
suffered severely during the E–O transition,
which brings support to the hypothesis that there
was significant sea-level fall associated with the
E–O climatic transition. Collection and correlation
of fossil records from these types of environments
may help further constrain the problem of identify-
ing and quantifying evidence for sea-level change
associated with the EOGM.
Multiple proxies support the idea that the biotic
changes were associated with major climatic and
oceanographic reorganization brought about by
widespread glaciation on Antarctica and possibly
in the Northern Hemisphere. The causal
THE E–O TRANSITION 375

mechanism for this fundamental shift is controver-
sial. The most popular explanations are that glacia-
tion was caused by (i) cooling of Antarctica due to
plate tectonic reorganization and related changes in
ocean circulation controlling poleward heat trans-
port and (ii) a threshold response to declining
atmospheric pCO
2
. Models and proxy records are
being interpreted very differently to bring support
to both theories. In our view, the balance is tipped
in favour of the idea that the Eocene–Oligocene
climate transition was forced by a gradual change
in a key parameter (i.e. pCO
2
) past a switching
point, rather than rapid forcing, which caused a
shift towards conditions favourable for ice
accumulation and leading to the geologically
sudden build-up of the Antarctic ice-cap. This
mechanism seems to provide the most parsimonious
explanation linking up proxy data observations,
orbital configurations and complex carbon system
and biosphere feedbacks. The effects were not
instantaneous, however, and were spread over a
timeframe of several hundred thousand years. On
the other hand, it is difficult to dismiss the tectonic
separation of Antarctica leading to geographic iso-
lation as pure coincidence. Most likely, the E –O
transition, which represents a very significant shift
in the Earth System, was caused by some complex
combination of forcing factors rather than a single
identifiable cause.
Incoming datasets have a tendency to raise
new, more focused questions that direct future
research. For example, d
18
O and Mg/Ca proxies
and the problems associated with separating temp-
erature and ice volume effects have peaked interest
in the possibility of contemporaneous Northern
Hemisphere glaciation. This highlights the neces-
sity for future deep-sea drilling to target the North-
ern Hemisphere, particularly Arctic, sediments that
will resolve this question. In addition, future drilling
should aim to sample sediments close to Antarctica
that might provide further constraints on the timing
and extent of Antarctic glaciation. Recognition of
the synchrony of CCD deepening and the d
18
O
shift into the EOGM has raised questions about
the importance of deep-ocean alkalinity changes
as a primary cause or effect of the climatic tran-
sition. Attention to these issues has also prompted
studies into the role of deep-water alkalinity and
carbonate ion concentration on the Mg/Ca palaeo-
temperature proxy (see Lear 2007).
In the future, production of high-resolution
records will be vital. First, these are essential for
resolving leads and lags between ice fluctuations
and carbon cycling. This has already been
achieved in one site (e.g. Coxall et al. 2005) but
needs to be duplicated elsewhere. Secondly, high-
resolution records on 1000-yr timescales are
necessary to assess the importance of orbital
pacing of climate, and identify changes in the influ-
ence of different forcing periods during the tran-
sition from a (relatively) ice-free to a glaciated
climate state. The Site 1218 record, which com-
bines various proxies, represents an important step
towards this goal and should provide a model for
future high-resolution studies.
In summary, the improved resolution of incom-
ing proxy records together with the variety of
additional proxies that provide independent checks
on key climatic parameters are providing better
constraints for modelling E–O climate change
scenarios. These data allow us to test hypotheses
on climate change mechanisms that incorporate
many factors. Model outputs force us to cross-check
data interpretations, seek new ways of constraining
key processes and, in some cases even re-evaluate
the principles underpinning proxies. In particular,
new SST estimates from ‘glassy forams’ and the
TEX
86
method will be important in this respect
and recent data for other time periods (Pearson
et al. unpublished) are already much closer to the
modelled view of warmer poles and tropics under
greenhouse conditions. Only then can the influence
of global extinction patterns on climate itself start to
be understood.
Still requiring further study is the question of the
relative importance of changing ocean circulation
patterns (as a consequence of the opening of
Southern Ocean tectonic gateways) and threshold-
global cooling related to gradually declining atmos-
pheric CO
2
. Resolving this issue requires improved
proxy records of atmospheric CO
2
at sufficient res-
olution to resolve leads and lags that reveal the
primary forcing mechanism. Future efforts should
be concentrated on refining the calibrations and
reducing assumptions to produce more accurate
pCO
2
estimates and obtaining higher-resolution
time series that can be accurately correlated with
glaciation and carbon cycle proxies, not forgetting
the power of multiple proxies to bring conciliation.
In addition, there is a great need to expand floral and
faunal datasets that document climatic and biotic
changes in the terrestrial realm, and to more accu-
rately correlate them with marine records. A funda-
mental hole in our understanding is the extent of
climate change and ice growth in the Arctic
region during the E–O transition. Efforts could be
focused on identifying and accessing records on
the Arctic margins and locating suitable deep-sea
drilling targets in the Arctic Ocean basin where
mid-Cenozoic sediments are preserved.
At the time of writing (August 2006), there is a
need for refinement of pCO
2
proxies, expansion of
datasets and improved understanding of high north-
ern latitude and continental E–O climate change.
Future application of new and existing proxies,
recovery of additional high-quality sediment
H. K. COXALL & P. N. PEARSON376

archives from the oceans and continents and more-
complex models that integrate chemical and bio-
logical processes across the Eocene–Oligocene
Transition will no doubt yield exciting results and
should rapidly advance the field of palaeoclimatol-
ogy as a whole.
References
ADA MS, C. G. 1965. The Foraminifera and stratigraphy of
the Melinau Limestone, Sarawak, and its importance
in Tertiary correlation. Quarterly Journal of the Geo-
logical Society of London,121, 283– 338.
ADA MS, C. G., BU TTE RLI N, J. & SA MANT A, B. K. 1986.
Larger foraminifera and events at the Eocene –
Oligocene boundary in the Indo-West Pacific. In:
POM EROL , C. & PRE MOLI SI LVA , I. (eds) Terminal
Eocene Events, Amsterdam, Elsevier, 237–252.
AUB RY, M.-P. 1991. Sequence stratigraphy: Eustacy or
tectonic imprint? Journal of Geophysical Research,
96, 6641– 6679.
AUB RY, M.-P. 1992. Late Paleogene nannoplankton evol-
ution: a tale of climatic deterioration. In: PR OTH ERO ,
D. R. & BE RGGR EN, W. A. (eds) Eocene –Oligocene
Climatic and Biotic Evolution, Princeton, NJ, Prince-
ton University Press.
AVE RY, K. B., CALHO UN, M., S CHM ALZ, L. S. &
PAY TAN, A. 2005. Data report: Carbonate and barite
trends across the Eocene–Oligocene Boundary at
Shatsky Rise, ODP Leg 198. In: BR ALO WER , T. J.,
PRE MOLI SI LVA , I. & MAL ONE, M. J. (eds) Proceed-
ings of the Ocean Drilling Program, Scientific Results,
College Station,TX, Ocean Drilling Program [Online].
Available from http://www-odp.tamu.edu/publi
cations/198_SR VOLUME/CHAPTERS/106. PDF,
1–16.
BAL DAUF , J. G. 1992. Middle Eocene through early
Miocene diatom floral turnover. In: PR OTHE RO,
D. R. & BE RGGR EN, W. A. (eds) Eocene –Oligocene
Climatic and Biotic Evolution, Princeton, NJ,
Princeton University Press.
BAL DAUF , J. G. & BAR RON , J. A. 1990. Evolution of bio-
siliceous sedimentation patterns – Eocene through
Quaternary: paleoceanographic response to polar
cooling. In: BL EIL, U. & T HIE DE, R. (eds) Geological
History of the Polar Oceans: Arctic versus Antarctic,
Norwell, MA, Kluwer Academic, 575–607.
BAR KER, P. F. & T HOMAS, E. 2004. Origin, signature and
palaeoclimatic influence of the antarctic circumpolar
current. Earth Science Reviews,66, 143– 162.
BAR KER, P. F., B ARR ETT , P. J., COO PER, A. K. &
HUYBRECHTS, P. 1999. Antarctic glacial history from
numerical models and continental margin sediments.
Palaeogeography, Palaeoclimatology, Palaeoecol-
ogy,150, 247– 267.
BAR NOLA , J.-M., RAY NAU D, D., KOR OTK EVIC H, Y. S.
& LO RIUS , C. 1987. Vostok ice core provides
160,000– year record of atmospheric CO
2
.Nature,
329, 408– 414.
BAR RERA , E. & HUB ER, B. T. 1991. Paleogene and
early Neogene oceanography of the southern Indian
Ocean: Leg 119 foramincfer stable isotope results.
Scientific Results of the Ocean Drilling Program,
119, 693– 717.
BAR RETT , P. J., HAM BRE Y, M. J., HAR WOO D, D. M.,
PYN E, A. R. & WEB B, P. N. 1989. Synthesis. In:
BAR RETT , P. J. (ed.) Antarctic Cenozoic History
from CIROS–1 Drill Hole, McMurdo Sound Antarc-
tica, Wellington, New Zealand, DSIR Publishing,
Department of Scientific and Industrial Research Bul-
letin, 245, 241– 251.
BEE RLIN G, D. J., CHALONER, W. G., H UNT LEY , B.,
PEA RSON , J. A. & TOO LEY , M. J. 1993. Stomatal
density responds to the glacial cycle of environmental
change. Proceedings of the Royal Society of London
Series B,251, 133– 138.
BEN SON, R. 1975. The origin of the psychrosphere as
recorded in changes of deep-sea ostracode assem-
blages. Lethaia,8, 69– 83.
BER GER, W. H. & W INT ERE R, E. L. 1974. Plate strati-
graphy and the fluctuating carbonate line. In: HS U
¨,
K. J. & JE NKY NS, H. C. (eds) Pelagic Sediments on
Land and Under the Sea, Special Publication of the
International Association of Sedimentologists, 11– 48.
BER GGRE N, W. A. & PE ARSO N, P. 2005. A revised tropi-
cal to subtropical Paleogene planktonic foraminiferal
zonation. Journal of Foraminiferal Research,35,
279–298.
BER GGRE N, W. A., KE NT, D. V., O BRA DOVI CH, J. D. &
SWI SHER , C. C. III. 1992. Toward a revised
Paleogene geochronology. In: PRO THE RO, D. R. &
BER GGRE N, W. A. (eds) Eocene –Oligocene Climatic
and Biotic Evolution, Princeton, NJ, Princeton Univer-
sity Press, 29– 45.
BER GGRE N, W. A., KE NT, D. V., SWIS HER, I. C. C. &
AUB RY, M. P. 1995. A Revised Cenozoic geochronol-
ogy and Chronostratigraphy. In: BE RGGR EN, W. A.,
KEN T, D. V. & HAR DEN BOL , J. (eds) Geochronology,
Time Scales and Global Stratigraphic Correlation: A
Unified Temporal Framework for an Historical
Geology, Society for Sedimentary Geology, Special
Publications, 129– 212.
BER NER, R. A. 1990. Atmospheric carbon dioxide levels
over Phanerozoic time. Science,249, 1382–1386.
BER NER, R. A. 1992. Palaeo-CO
2
and climate; discussion.
Nature,358, 114.
BER NER, R. A. 1997. The carbon cycle and CO
2
over Pha-
nerozoic time: the role of land plants. Philosophical
Transactions of the Royal Society Series B,353,
75–82.
BER NER, R. A. & KOTHA VAL A, Z. 2001. GEOCARB III.
A revised model of atmospheric CO
2
over Phanerozoic
time. American Journal of Science,291, 182–204.
BIL LUPS , K. & SCH RAG , D. P. 2003. Application of
benthic foraminiferal Mg/Ca ratios to questions of
Cenozoic climate change. Earth and Planetary
Science Letters,209, 181– 195.
BIR KENM AJE R, K. 1996. Tertiary glacial/interglacial
palaeoenvironments and sea-level changes, King
George Island, West Antarctica. An overview. Bulletin
of the Polish Academy of Sciences, Earth Sciences,44,
157–181.
BIR KENM AJE R, K., GA ZDZI CKI , A., KRA JEWS KI, K. P.,
PRZ YBYC IN, A., S OLE CKI, A. & T ATU R, YO ON, H.
II. 2005. First Cenozoic glaciers in West Antarctica.
Polish Polar Research,26, 3–12.
THE E–O TRANSITION 377

BLO W, W. H. 1979. The Cainozoic globigerinida. A study
of the morphology, taxonomy, evolutionary relation-
ships and the stratigraphical distribution of some of
the Globigerinida (mainly Globigerinacea). Leiden,
E. J. Brill,3, 1413pp.
BLO W, W. H. & BAN NER , F. T. 1962. The mid-Tertiary
(upper Eocene to Aquitanian) Globigerinacae. In:
EAM ES, F. T. ET A L. (eds). Fundamentals of mid-
Tertiary Stratigraphical Correlation, London,
Cambridge University Press, 61–151.
BOE RSMA , A. 1986. Eocene– Oligocene Atlantic paleo–
oceanography, using benthic foraminifera. In:
POM EROL , C. & PRE MOLI SI LVA , I. (eds) Terminal
Eocene Events, Elsevier, 225–236.
BOE RSMA , A. & PRE MOL I SILVA , I., 1986. Terminal
Eocene events – planktonic foraminifera and isotopic
evidence. In: PO MERO L, C. & PR EMO LI SILV A, I.
(eds) Terminal Eocene Events, Elsevier, 213–224.
BOT TOML EY, R., GRIE VE, R., YO RK, D. & M ASA LTIS ,
V. 1997. The age of the Popigai impact event and its
relation to events at the Eocene –Oligocene boundary.
Nature,388, 365– 368.
BOU COT, A. J. & G RAY , J. 2001. A critique of Phanero-
zoic climatic models involving changes in the CO
2
content of the atmosphere. Earth–Science Reviews,
56, 1–159.
BRE ZA, J. R. & WI SE, S. W. J R. 1992. Lower Oligocene
ice– rafted debris on the Kerguelen Plateau: evidence
for East Antarctic continental glaciation. In: WIS E, S.
W. & SC HLIC H, R. ET A L. (eds) Proceedings of the
Ocean Drilling Program Scientific Result, vol. 120,
College Station, TX, Ocean Drilling Program, 161– 178.
BRY ANT, H. N. 1996. Nimravidae. In: P ROT HERO , D. R.
& EM RY, R. J. (eds) The Terrestrial Eocene –Oligo-
cene Transition in North America, Cambridge Univer-
sity Press, 453–475.
BRI NKHU IS, H. 1992. Late Paleogene Dinoflagellate
Cysts with special reference to the Eocene–Oligocene
boundary. In: PRO THE RO, D. R. & B ERG GREN , W. A.
(eds) Eocene– Oligocene Climatic and Biotic Evol-
ution, Princeton, NJ, Princeton University Press,
327–340.
BRI NKHU IS, H. & B IFF I, U. 1993. Dinoflagellate cysts
stratigraphy of the Eocene –Oligocene transition in
central Italy. Marine Micropalaeontology,22,
131–183.
BUR KE, W. H., DENIS ON, R. E., HEATH ERI NGT ON, E.
A., KO EPNI CK, R. B., NELS ON, H. F. & OTTO, J. B.
1982. Variation of seawater
87
Sr/
86
Sr throughout Pha-
nerozoic time. Geology,10, 516– 519.
CAN DE, S. C. & KE NT, D. V. 1995. Revised calibration of
the geomagnetic polarity timescale for the Late Cre-
taceous and Cenozoic. Journal of Geophysical
Research,100, 6093– 6095.
CAR TER, B. D. 2003. Diversity patterns in Eocene and
Oligocene echinoids of the Southeastern United
States. In: PR OTHE RO, D. R., I VAN Y, L. &
NES BITT , E. A. (eds) From Greenhouse to Icehouse,
New York, Columbia University Press, 354 –365.
CER LING , T. E. 1991. Carbon dioxide in the atmosphere:
evidence from Cenozoic and Mesozoic paleosols.
American Journal of Science,291, 377–400.
CER LING , T. E. 1999. Stable carbon isotopes in paleosol
carbonates: evidence from Cenozoic and Mesozoic
paleosols. Special Publication of the International
Association of Sedimentologists,27, 43– 60.
CHA MLEY , H. 1986. Clay mineralogy at the Eocene–
Oligocene boundary. In: PO MERO L, C. & PR EMO LI
SIL VA, I. (eds) Terminal Eocene Events, Elsevier,
381–386.
COC CION I, R. & GA LEO TTI, S. 2003. Deep water benthic
foraminiferal events from the Massignano Eocene–
Oligocene boundary stratotype, Central Italy. In:
PRO THER O, D. R., IV ANY , L. & NES BITT , E. A.
(eds) From Greenhouse to Icehouse, Columbia
University Press.
COC CION I, R., MO NAC O, P., MON ECH I, S., NO CCHI , M.
& PA RISI , G. 1988. Biostratigraphy of the Eocene–
Oligocene Boundary at Massignano (Ancona, Italy).
International Subcommission on Paleogene Strati-
graphy, E–O Meeting, 59 –75.
COL LINS ON, M. E. 1992. Vegetation and floristic changes
around the Eocene–Oligocene boundary in western
and central Europe. In: PR OTHE RO, D. R. & B ERG G-
RE N, W. A. (eds) Eocene–Oligocene Climatic and
Biotic Evolution, Princeton, NJ, Princeton University
Press.
COR SINI , J., SMI TH, T. & LEITE , M. 2006. Palaeoenvir-
onmental implications of size, carapace position,
and incidence of non-shell elements in White
River turtles. Palaeogeography, Palaeoclimatology,
Palaeoecology,234, 287– 303.
COX ALL, H. K. & P EAR SON , P. N. 2006. Taxonomy,
biostratigraphy and phylogeny of Hantkeninidae
(Clavigerinella, Hantkenina and Cribrohantkenina).
In: PEA RSO N, P. N., OLS SON, R. K., H EML EBEN , C.,
HUB ER, B. T. & BE RGG REN , W. A. (eds) Atlas
of Eocene Planktonic Foraminifera, Cushman
Foundation of Foraminiferal Research, Special
Publication No. 41, Lawrence, Kansas, Allen Press,
213– 252.
COX ALL, H. K., WI LSO N, P. A., PA
¨LI KE, H., LE AR, C. H.
& BA CKMA N, J. 2005. Rapid stepwise onset of Ant-
arctic glaciation and deeper calcite compensation in
the Pacific Ocean. Nature,433, 53– 57.
CRO WLEY , T. J. 2000. Carbon dioxide and Phanerozoic
climate: An overview. In: HU BER, B. T.,
MAC LEOD, K. G. & W ING , S. L. (eds) Warm Cli-
mates in Earth History, Cambridge University Press,
425–444.
CUR RY, W. B., SHACK LET ON, N. J., RICHT ER, C. E T AL.
1995. Proceedings of the Ocean Drilling Program,
Initial Reports, 154: College Station,TX, Ocean Dril-
ling Program, 1111pp.
DAL L’AN TON IA, B., BOSSI O, A. & GU ERN ET, C. 2003.
The Eocene–Oligocene boundary and the psychro-
spheric event in the Tethys as recorded by deep-sea
ostracodes from the Massignano Global Boundary
Stratotype Section and Point, Central Italy. Marine
Micropaleontology,48, 91– 106.
DAL AI, T. K., RAVIZ ZA, G. E. & PEUCK ER – E HREN -
BR INK, B. 2006. The Late Eocene
187
Os/
188
Os excur-
sion: Chemostratigraphy, cosmic dust flux and the
Early Oligocene glaciation. Earth and Planetary
Science Letters,241, 477– 492.
DAV IES, R., C ART WRI GHT, J., P IKE , J. & LIN E, C. 2001.
Early Oligocene initiation of North Atlantic deep
water formation. Nature,410, 917– 920.
H. K. COXALL & P. N. PEARSON378

DECON TO, R. M. & PO LLA RD, D. 2003a. Rapid Ceno-
zoic glaciation of Antarctica induced by declining
atmospheric CO
2
.Nature,421, 245– 249.
DECON TO, R. M. & PO LLA RD, D. 2003b. A coupled
climate-ice sheet modeling approach to the Early
Cenozoic history of the Antarctic ice sheet. Palaeo-
geography, Palaeoclimatology, Palaeoecology,198,
39– 52.
DEC ONTO , R. & POL LAR D, D. 2007. Rethinking Ceno-
zoic glacial history: a model-data perspective.
Geophysical Research Abstracts,9, 1607-7962/gra/
EGU2007-A-09083.
DIEKMANN, B., KU HN, G., GERSO NDE , R. & MAC KEN-
SE N, A. 2004. Middle Eocene to early Miocene
environmental changes in the sub-Antarctic Southern
Ocean: evidence from biogenic and terrigenous
depositional patterns at ODP Site 1090. Global and
Planetary Change,40, 295– 313.
DIE STER -HA ASS, L. 1992. Late Eocene –Oligocene sedi-
mentation in the Antarctic Ocean, Atlantic sector
(Maud Rise, ODP Leg 113, Site 689): development
of surface and bottom water circulation. Antarctic
Research Series,56, 185–202.
DIE STER -HA ASS, L. 1995. Middle Eocene to early Oligo-
cene palaeoceanography of the Antarctic Ocean
(Maud Rise, ODP Leg 113. Site 689): change from a
low productivity to a high productivity ocean. Palaeo-
geography, Palaeoclimatology, Palaeoecology,113,
311–334.
DIE STER -HA ASS, L. 1996. Late Eocene –Oligocene
paleoceanography in the southern Indian Ocean
(ODP Site 744). Marine Geology,130, 99–119.
DIE STER -HA ASS, L. & ZA CHOS , J. C. 2003. The
Eocene– Oligocene transition in the Equatorial Atlan-
tic (ODP Site 925): Paleoproductivity increase and
positive d
13
C excursion. In: PRO THE RO, D. R.,
IVA NY, L. & NESBI TT, E. A. (eds) From Greenhouse
to Icehouse, New York, Columbia University Press,
397–418.
DIE STER -HA ASS, L. & Z AHN , R. 1996. Eocene–
Oligocene transition in the Southern Ocean; history
of water mass circulation and biological productivity.
Geology,24, 163– 166.
DIE STER -HA ASS, L. & Z AHN , R. 2001. Palaeopro-
ductivity increase at the Eocene –Oligocene climatic
transition; ODP/DSDP sites 763 and 592. Palaeo-
geography, Palaeoclimatology, Palaeoecology,172,
153–170.
DIE STER -HA ASS, L., R OBER T, C. & CH AML EY, H. 1993.
Paleoceanographic and paleoclimatic evolution in the
Weddell Sea (Antarctica) during the middle Eocene –
late Oligocene, from a coarse sediment fraction and
clay mineral data (ODP Site 689). Marine Geology,
114, 233– 250.
DOC KERY , D. T. III. 1986. Punctuated succession of
marine molluscs in the northern Gulf Coastal Plain.
Palaios,1, 582– 589.
DOC KERY , D. T. III . & LOZ OUET , P. 2003. Biotic patterns
in Eocene–Oligocene Molluscs of the Atlantic Coastal
Plain, USA. In: PR OTHE RO, D. R., I VAN Y, L. &
NES BITT , E. A. (eds) From Greenhouse to Icehouse,
New York, Columbia University Press, 303–340.
DUPONT-NIVET , G., KRI JGSM AN, W., L ANG EREI S, C. G.,
ABE LS, H. A. & FA NG, X. 2007. Tibetan plateau
aridification linked to global cooling at the Eocene-
Oligocene transition. Nature,445, 635– 638.
EAG LES, G., L IVE RMOR E, R. & MO RRI S, P. 2006. Small
basins in the Scotia Sea: The Eocene Drake Passage
gateway. Earth and Planetary Science Letters,242,
343–353.
ECH OLS, R., A RME NTR OUT, J. M., R OOT , S. A., FEA RN,
L. B., CO OKE, J. C., R ODG ERS, B. K. & T HOM PSO N,
P. R. 2003. Sequence stratigraphy of the Eocene–
Oligocene Boundary Interval: Southeastern Missis-
sippi. In: PR OTHE RO, D. R., I VAN Y, L. & NES BIT T,
E. A. (eds) From Greenhouse to Icehouse,
New York, Columbia University Press, 189–222.
EHR MANN , W. U. & MAC KEN SEN, A. 1992. Sedimento-
logical evidence for the formation of an East Antarctic
ice sheet in Eocene –Oligocene time. Palaeogeo-
graphy, Palaeoclimatology, Palaeocecology,93,
85–112.
EKA RT, D. D., CE RLI NG, T. E., MONTA N
˜EZ , I. P. &
TAB OR, N. J. 1999. A 400 million year carbon
isotope record of pedogenic carbonate: implications
for palaeoatmospheric carbon dioxide. American
Journal of Science,299, 805–827.
ELD RETT , J. S., HAR DIN G, I. C., WIL SON , P. A.,
BUT LER, E. & R OBE RTS , A. P. 2007. Continental
ice in Greenland during the Eocene and Oligocene.
Nature, doi:10.1038/nature05591
EXON, N. F., KE NNET T, J. P. & MA LON E, M. J. 2004.
Leg 189 Synthesis: Cretaceous –Holocene history of
the Tasmanian gateway. In: EXON, N. F., KE NNE TT,
J. P. & MA LONE , M. J. ET AL . (eds) Proceedings of
the Ocean Drilling Program: Scientific Results, 189.
[Online]. Available from http://www-odp://www.odp.
tamu. edu/publications/189_SR/synth/synth. htm.
FLU EGEM AN, R. H. 2003. Late Eocene –early Oligocene
in the Gulf Coastal Plain: regional vs. global influ-
ences. In: PR OTHE RO, D. R., I VAN Y, L. &
NES BITT , E. A. (eds) From Greenhouse to Icehouse,
New York, Columbia University Press, 283–293.
FRA NCIS , J. E. 1999. Evidence from fossil plants for
Antarctica palaeoclimates over the past 100 million
years, Terra Antarctica,3, 43– 52.
FRA NCIS , J. E. & POO LE, I. 2002. Cretaceous and Tertiary
climates of Antarctica: evidence from fossil wood.
Palaeogeography, Palaeoclimatology, Palaeo-
ecology,182, 47– 64.
FOR DYCE , R. E. 1992. Cetacean evolution and Eocene –
Oligocene environments. In: PRO THE RO, D. R. &
BER GGRE N, W. A. (eds) Eocene –Oligocene Climatic
and Biotic Evolution, Princeton, Princeton University
Press, 368–382.
FOR DYCE , R. E. 2003. Cetacean evolution and Eocene –
Oligocene oceans revisited. In: PR OTHE RO, D. R.,
IVA NY, L. & NE SBI TT, E. A. (eds) From Greenhouse
to Icehouse, New York, Columbia University Press,
368–382.
FUNAKAWA, S., NI SHI, H., MOOR E, T. C. & NIG RIN I,
C. A. 2006. Radiolarian faunal turnover and palaeocea-
nographic change around the Eocene–Oligocene
boundary in the Central Equatorial Pacific, ODP Leg
199, Holes 1218A, 1219A and 1220A. Palaeogeogra-
phy, Palaeoclimatology, Palaeoecology,230, 183– 203.
GAL E, A. S., HUG GET T, J. M., PA
¨LI KE, H., LA URI E, E.,
HAL IWOO D, E. A. & HA RDE BOL, J. 2006. Correlation
THE E–O TRANSITION 379

of Eocene–Oligocene marine and continental records:
orbital cyclicity, magnetostratigraphy and sequence
stratigraphy of the Solent Group, Isle of Wight, UK.
Journal of the Geological Society of London,163,
401–415.
GED L, P. 2004. Dinoflagellate cyst record of the Eocene–
Oligocene boundary succession in flysch deposits at
Lelucho
´w, Carpathian Mountains, Poland. Geological
Society Special Publication,230, 309– 324.
GIL BERT , R., DOM ACK , E. W. & CAME RLE NGHI , A.
2003. Deglacial history of the Greenpeace
Trough: Icesheet to ice shelf transition in the north-
western Weddell Sea. Antarctic Research Series,79,
195–204.
GOL DSTE IN, S. L. & JACOB SEN , S. B. 1988. Nd and Sr
isotope systematics of river water suspended material:
Implications for crustal evolution. Earth and Plane-
tary Science Letters,87, 249–265.
GRI MES, S. T., HOO KER , J. J., COL LINS ON, M. E. &
MAT TEY, D. P. 2005. Summer temperatures of late
Eocene to early Oligocene freshwaters. Geology,33,
189–192.
HAQ , B. U., HARD ENB OL, J. & VEIL , P. U. 1987. Chron-
ology of fluctuating sea levels since the Triassic.
Science,235, 1156– 1166.
HAR RISO N, K. G. 2000. Role of increased marine silica
input on paleo-pCO
2
levels. Paleoceanography,13,
292–298.
HAR TL, P., TA UXE , L. & HER BERT , T. 1995. Earliest Oli-
gocene increase in South Atlantic productivity as inter-
preted from “rock magnetics” at Deep-sea Drilling
Project Site 522. Paleoceanography,10, 311– 326.
HAY , W., FLO
¨GE L, S. & SO
¨DI NG, E. 2005. Is the
initiation of glaciation on Antarctica related to
a change in the structure of the ocean? Global and
Planetary Change,45, 23–33.
HES S, J., BEN DER , M. L. & SCH ILLI NG, J.-G. 1986. Evol-
ution of the ratio of strontium-87 to strontium-86 in
seawater from Cretaceous to Present. Science,231,
979–984.
HIC KMAN , C. S. 2003. Evidence for Abrupt Eocene–
Oligocene molluscan faunal change in the Pacific
Northwest. In: PR OTHE RO, D. R., I VAN Y, L. &
NES BITT , E. A. (eds) From Greenhouse to Icehouse,
New York, Columbia University Press, 71 –87.
HOO KER, J. J., C OLL INS ON, M. E. & SI LLE , N. P. 2004.
Eocene–Oligocene mammalian faunal turnover in the
Hampshire Basin, UK; calibration to the global time
scale and the major cooling event. Journal of the Geo-
logical Society of London,161, 161– 172.
HO
¨NI SCH, B. & H EMM ING , N. G. 2004. Ground-truthing
the boron isotope-paleo-pH proxy in planktonic forami-
nifera shells: Partial dissolution and shell size effects.
Paleoceanography,19, doi:10.1029/2004PA001026.
HSU
¨, J. K., LABR EQU E, J. J., CAR MAN , M. F. ET AL .
1984. Initial Reports of the Deep-Sea Drilling
Project,73: Washington DC, US Government Printing
Office, 798pp.
HUB ER, M., BR INK HUI S, H., STI CKL EY, C. E. E T AL.
2004. Eocene circulation of the Southern Ocean:
Was Antarctica kept warm by subtropical waters?
Paleoceanography,19, doi:10.1029/2004PA001014.
HUT CHIN SON , J. H. 1992. Western North American
reptile and amphibian record across the Eocene –
Oligocene boundary and its climatic implications. In:
PRO THER O, D. R. & BE RGG REN, W. A. (eds)
Eocene– Oligocene Climatic and Biotic Evolution,
Princeton, NJ, Princeton University Press, 451–463.
IMB RIE, J., H AYS, J. D., MCINTY RE, A. E T AL. 1984. The
orbital theory of Pleistocene climate: Support from a
revised chronology of the marine d
18
O record. In:
BER GER, A., I MBR IE, J., HAYS, J., KU KLA , G. J. &
SAL TZMA N, E. (eds) Milankovitch and Climate,
Boston, Reidel, 269–305.
IVA NY, L. C., PATTE RSO N, W. P. & LOH MAN N, K. C.
2000. Cooler winters as a possible cause of mass
extinctions at the Eocene–Oligocene boundary.
Nature,407, 887– 890.
IVA NY, L., NE SBI TT, E. A. & PR OTH ERO , D. R. 2003.
The Marine Eocene–Oligocene transition – A Syn-
thesis. In: PR OTHE RO, D. R., I VAN Y, L. &
NES BITT , E. A. (eds) From Greenhouse to Icehouse,
New York, Columbia University Press, 522–534.
IVA NY, L. C., VA N SIMA EYS , S., DOMA CK, E. W. &
SAM SON, S. D. 2006. Evidence for an earliest Oligo-
cene ice sheet on the Antarctic Peninsula. Geology,
34, 377– 380.
JAR AMIL LO, C., R UED A, M. J. & MOR A, G. 2006. Ceno-
zoic plant diversity in the Neotropics. Science,311,
1893– 1896.
JOV ANE, L., F LOR IND O, F. & DIN ARE
`S-TUREL L, J.
2004. Environmental magnetic record of paleoclimate
change from the Eocene–Oligocene stratotype
section, Massignano, Italy. Geophysical Research
Letters,31, doi:10.1029/2004GL020554.
JOV ANE, L., F LOR IND O, F., SPR OVI ERI, M. & PA
¨LI KE,
H. 2006. Astronomic calibration of the late
Eocene/early Oligocene Massignano section (central
Italy). Geochemistry, Geophysics, Geosystems,7,
doi:10.1029/2005GC001195.
KAM INSK I, M. A. 2005. The utility of Deep-Water
Agglutinated Foraminiferal acmes for correlating
Eocene to Oligocene abyssal sediments in the North
Atlantic and Western Tethys. Studia Geologica Polo-
nica. 124, 325– 339.
KELLER, G. 1983. Biochronology and paleoclimatic impli-
cations of middle Eocene to Oligocene planktonic for-
aminiferal faunas. Micropalaeontolgy,7, 463– 486.
KELLER, G. 1986. Stepwise mass extinctions and impact
events: Late Eocene to early Oligocene. Marine
Micorpaleontology,10.
KELLER, G., MAC LEO D, N. & BA RRER A, E. 1992.
Eocene– Oligocene faunal turnover in planktonic fora-
minifera, and Antarctic glaciation. In: PR OTH ERO, D.
R. & BE RGGR EN, W. A. (eds) Eocene and Oligocene
Climatic and Biotic Evolution, Princeton, Princeton
University Press, 218–244.
KEM P, E. M. 1978. Tertiary climatic evolution and veg-
etation history in the southeast Indian Ocean Region.
Palaeogeography, Palaeoclimatology, Palaeo-
ecology,24, 169– 208.
KEN NETT , J. P. 1977. Cenozoic evolution of Antarctic
glaciation, the circum-Antarctic Ocean, and their
impact on global paleoceanogrpahy. Journal of Geo-
physical Research,82, 384– 386.
KEN NETT , J. P. & SHA CKL ETON , N. J. 1976. Oxygen iso-
topic evidence for the development of the psychro-
sphere 38 Myr ago. Nature,260, 513– 515.
H. K. COXALL & P. N. PEARSON380

KEN NETT , J. P. & ST OTT, L. D. 1990. Proteus and
Proto-Oceanus: ancestral Paleogene oceans as revealed
from stable isotopic results: ODP Leg 113. In:
BAR KER, P. F. & KE NNE TT, J. P. E T AL. (eds) Proc-
eedings of the Ocean Drilling Program, Scientific
Results, College Station, TX, Ocean Drilling Program,
937– 960.
KEN NETT , J. P., HOU TZ, R. E., A NDR EWS, P. B. &
EDW ARDS , A. R. 1975. Development of the circum-
Antarctic current. Science,186, 144– 147.
KID D, R. B. & HIL L, P. R. 1987. Sedimentation on Feni
and Gardar sediment drifts. Initial Reports of the
Deep-Sea Drilling Project,94, 1217–1244.
KIE SSLI NG, W., F LU
¨GE L, E. & GOL ONK A, J. 2003. Pat-
terns of Phanerozoic carbonate platform sedimen-
tation. Lethaia,36, 195– 225.
KOB ASHI , T., GRO SSM AN, E. L., YA NCE Y, T. E. &
DOC KERY , D. T. III . 2001. Reevaluation of conflict-
ing Eocene tropical temperature estimates: Molluscan
oxygen isotope evidence for warm low latitudes.
Geology,29, 983– 986.
KUM P, L. R., BRA NTL EY, S. L. & ARTHU R, M. A. 2000.
Chemical weathering, atmospheric CO
2
, and climate.
Annual Review of Earth and Planetary Sciences,28,
611–667.
LAS KAR, J. 1999. The limits of Earth orbital calculations
for geological time-scale use. Royal Society Philoso-
phical Transactions: Mathematical, Physical and
Engineering Sciences (Series A),193, 1735– 1759.
LAS KAR, J., R OBU TEL , P., JOUT EL, F., GAST INEA U, M.,
COR REIA , A. & LEV RAR D, B. 2004. A long term
numerical solution for the insolation quantities of the
Earth. Astronomy and Astrophysics,428, 261– 285.
LAT IMER , J. & FIL IPPE LLI , G. 2002. Eocene to Miocene
terrigenous inputs and export production: geochemical
evidence from ODP Leg 177, Site 1090. Palaeogeo-
graphy, Palaeoclimatology, Palaeocecology,182,
151–164.
LAW VER, L. A. & GA HAG AN, L. M. 2003. Evolution of
Cenozoic seaways in the circum-Antarctic region.
Palaeogeography, Palaeoclimatology, Palaeocecol-
ogy,198, 11– 37.
LEA R, C. H. 2007. Mg/Ca Palaeothermometry: A New
Windowinto Cenozoic ClimateChange. In: WI LLIA MS,
M., HAY WOO D, A. M., GREG ORY , F. J. &
SCH MIDT , D. N. (eds) Deep-time Perspectives on
Climate Change: Marrying the Signal from Computer
Models and Biological Proxies. The Micropalaeontolo-
gical Society, Special Publications. The Geological
Society, London, 313–322.
LEA R, C. H., EL DERF IEL D, H. & WIL SON , P. A. 2001.
Cenozoic deep-sea temperatures and global ice
volumes from Mg/Ca in benthic foraminiferal
calcite. Science,287, 269– 272.
LEA R, C. H., EL DERF IEL D, H. & WIL SON , P. A. 2003. A
Cenozoic sea water Sr/Ca record from benthic forami-
niferal calcite and its application in determining global
weathering fluxes. Earth and Planetary Science
Letters,208, 69– 84.
LEA R, C. H., RO SENT HAL , Y., COXA LL, H. K. &
WIL SON, P. A. 2004. Late Eocene to early Miocene
ice-sheet dynamics and the global carbon cycle.
Paleoceanography,19, PA4015, doi:10.1029/
2004PA001039.
LEO POLD , E. B., LIU , G. & CLA Y – POO LE, S. 1992.
Low–biomass vegetation in the Oligocene? In:
PRO THER O, D. R. & BE RGG REN, W. A. (eds)
Eocene and Oligocene Climatic and Biotic Evolution,
Princeton, Princeton University Press, 399–420.
LIU , Y. -J., AREN S, N. C. & LI, C.-S. 2007. Range change
in Metasequoia: Relationship to palaeoclimate. Bota-
nical Journal of the Linnean Society,154, 115–127.
LIU , Z., TOU, S., ZH AO, Q., CHEN G, X. & HUA NG, W.
2004. Deep-water Earliest Oligocene Glacial
Maximum (EOGM) in the South Atlantic. Chinese
Science Bulletin,49, 2190– 219.
LIV ERMO RE, R., N ANK IVEL L, A., EA GLE S, G. &
MOR RIS, P. 2005. Paleogene opening of Drake
Passage. Earth and Planetary Science Letters,236,
459–470.
LOCKWOOD, R. 2005. Body size, extinction events,
and the early Cenozoic record of veneroid bivalves:
A new role for recoveries? Paleobiology,31,
578–590.
LOU TRE, M. F., P AIL LAR D, D., VIM EUX , F. & COR TIJ O,
E. 2004. Does mean annual insolation have the poten-
tial to change the climate? Earth and Planetary
Science Letters,221, 1– 14.
LOW ENST EIN , T. K. & DEM ICCO , R. V. 2006. Elevated
Eocene atmospheric CO
2
and its subsequent decline.
Science,313, 1928.
MAN NING , E. M. 2003. The Eocene– Oligocene transition
in marine vertebrates of the Gulf Coastal Plain. In:
PRO THER O, D. R., IV ANY , L. & NES BITT , E. A.
(eds) From Greenhouse to Icehouse, New York,
Columbia University Press, 386–398.
MAR INO, L., M CSHEA , D. W. & UHEN , M. D. 2004.
Origin and evolution of large brains in toothed
whales. Anatomical Record – Part A Discoveries in
Molecular, Cellular, and Evolutionary Biology,281,
1247– 1255.
MAR SHAL L, L. G. & CI FELL I, R. L. 1989. Analysis of
changing diversity patterns in Cenozoic land
mammal age faunas, South America. Palaeoverteb-
rata,19, 169– 210.
MAR TIN, J. H. 1990. Glacial– interglacial CO
2
change:
the iron hypothesis. Paleoceanography,5, 1– 13.
MAR TIN, P. A., L EA, D. W., R OSE NTHA L, Y.,
SHA CKLE TON , N. J., SAR NTHE IN, M. & P APE NFUS S,
T. 2002. Quaternary deep-sea temperature histories
derived from benthic foraminiferal Mg/Ca. Earth
and Planetary Science Letters,189, 193– 209.
MAS HIOT TA, T. A., L EA, D. W. & SP ERO , H. J. 1999.
Glacial–interglacial changes in Subantarctic sea
surface temperature and d
18
O-water using foraminif-
eral Mg. Earth and Planetary Science Letters,170,
417–432.
MAY R, G. 2005. The Paleogene fossil record of birds in
Europe. Biological Reviews of the Cambridge Philoso-
phical Society,80, 515– 542.
MCGOW RAN, B. & B EEC ROF T, A. 1986. Neritic,
southern extratropical foraminifera and the terminal
Eocene event. Palaeogeography, Palaeoclimatology,
Palaeoecology,55, 23– 34.
MCGOW RAN, B., M OSS , G. & BEE CROF T, A. 1992.
Late Eocene and early Oligocene in Southern
Australia: Local neritic signals of global oceanic
changes. In: PRO THE RO, D. R. & BERGG REN ,
THE E–O TRANSITION 381

W. A. (eds) Eocene–Oligocene Climatic and Biotic
Evolution, Princeton, NJ, Princeton University Press,
178–201.
MEA D, G. A. & HOD ELL , D. A. 1995. Controls on the
87
Sr/
86
Sr composition of seawater from the middle
Eocene to Oligocene: Hole 689B, Maud Rise, Antarc-
tica. Paleoceanography,10, 327– 346.
MEA D, G. A., TA UXE, L. & L ABREQ UE, J. J. 1986. Oli-
gocene paleoceanography of the South Atlantic: paleo-
climatic implications of sediment accumulation rates
and magnetic susceptibility measurements. Paleocea-
nography,1, 273– 284.
MEN G, J. & MCKEN NA, M. C. 1998. Faunal turnovers of
Paleogene mammals from the Mongolian Plateau.
Nature,394, 364– 367.
MIL LER, K. G. & F AIR BAN KS, R. G. 1985. Oligocene to
Miocene carbon isotope cycles and abyssal circulation
changes. In: SU NQUI ST, E. T. & BROEC KER , W. S.
(eds) The Carbon Cycle and Atmospheric CO
2
;
Natural Variations Archean to Present, American
Geophysical Union, Geophysics Monograph 32,
469–486.
MIL LER, K. G. & M OUN TAI N, G. S. 1996. Drilling and
dating New Jersey Oligocene–Miocene sequences:
Ice volume, global sea level, and Exxon records.
Science,271, 1092– 1095.
MIL LER, K. G., FAI RBA NKS, R. & M OUN TAIN , G. S.
1987. Tertiary Oxygen isotope synthesis, sea level
history and continental margin erosion. Paleoceano-
graphy,1, 1–19.
MIL LER, K. G., WRI GHT , J. & FAIR BAN KS, R. 1991.
Unlocking the icehouse: Oligocene–Micoene oxygen
isotopes, eustacy and margin erosion. Journal of Geo-
physical Research,96, 6829– 6848.
MOH R, B. 1990. Eocene and Oligocene sporomorphs and
dinoflagellate cysts from Leg 113 drill sites, Weddell
Sea, Antarctica. In: BAR KER, P. F. & KE NNET T, J. P.
ET A L. (eds) Proceedings of the Ocean Drilling
Program, Scientific Results, College Station, TX,
Ocean Drilling Program, 595–612.
MOL INA, E. 1986. Description and biostratigraphy of the
main reference section of the Eocene–Oligocene
boundary in Spain: Fuente Caldera section. In:
POM EROL , C. & PRE MOLI SI LVA , I. (eds) Terminal
Eocene Events, Elsevier, 53–64.
MOL INA, E., G ONZ ALV O, C., ORT IZ, S. & CRUZ , L. E.
2006. Foraminiferal turnover across the Eocene–
Oligocene transition at Fuente Caldera, southern
Spain: No cause-effect relationship between meteorite
impacts and extinctions. Marine Micropaleontology,
58, 270–286.
MOO RE, M. J. & JA NSE N, R. K. 2005. Molecular evidence
for the age, origin, and evolutionary history of the
American desert plant genus Tiquilia (Boraginaceae).
Molecular Phylogenetics and Evolution,39, 668– 687.
MOO RE, T. C. J., RABIN OWI TZ, P. D. E T AL. 1984. Initial
Reports of the Deep-Sea Drilling Project,74,
Washington, DC, US Government Printing Office.
MOR AN, K., BA CKM AN, J. & BRINK HUI S, H. ET A L.
2006. The Cenozoic palaeoenvironment of the Arctic
Ocean. Nature,441, 601– 605.
MUR PHY, M. G. & K ENNE TT, J. P. 1986. Development of
latitudinal thermal gradients during the Oligocene:
Oxygen isotope evidence from the southwest Pacific.
Initial Reports of the Deep-Sea Drilling Project,
Washington DC, US Government Printing Office,
1347– 1360.
MYE RS, J. A. 2003. Terrestrial Eocene –Oligocene veg-
etation and climate in the Pacific Northwest. In:
PRO THER O, D. R., IV ANY , L. & NES BITT , E. A.
(eds) From Greenhouse to Icehouse, New York,
Columbia University Press, 171–188.
NEB ELSI CK, J. H., R ASS ER, M. W. & BASSI , D. 2005.
Facies dynamics in Eocene to Oligocene circumalpine
carbonates. Facies,51, 197– 216.
NES BITT , E. A. 2003. Changes in shallow marine faunas
from the Northeastern Pacific margin across the
Eocene– Oligocene boundary. In: PR OTHE RO, D. R.,
IVA NY, L. & NE SBI TT, E. A. (eds) From Greenhouse
to Icehouse, New York, Columbia University Press,
57–70.
NIL SEN, E. B., A NDE RSO N, J. B. & DEL ANE Y, M. L.
2003. Paleoproductivity, nutrient burial, climate
change and the carbon cycle in the western equatorial
Atlantic across the Eocene–Oligocene boundary.
Paleoceanography,18, doi:10.1029/2002PA000804.
NOC CHI, M., PAR ISI , G., MON ACO, P. ET A L. 1986. The
Eocene– Oligocene boundary event in the Umbrian
pelagic sequence, Italy. In: PO MERO L, C. &
PRE MOLI SI LVA , I. (eds) Terminal Eocene Events,
Amsterdam, Elsevier, 25–40.
NOC CHI, M., MO NEC HI, S., CO CCI ONI , R. ET AL . 1988.
The extinction of Hantkeninidae as a marker for defin-
ing the Eocene–Oligocene boundary: A proposal. In:
PRE MOLI SI LVA , I., COC CION I, R. & MO NTA NARI ,
A. (eds) International Union of Geological Sciences
Commission on Stratigraphy, The International Sub-
commission of Paleogene Stratigraphy: The Eocene–
Oligocene Boundary in the Marche –Umbria Basin
(Italy), Ancona, Monte Co
`nero, 249–252.
NON G, G. T., NA JJAR , R. G., SEI DOV , D. & PET ERSO N,
W. 2000. Simulation of ocean temperature change due
to the opening of Drake Passage. Geophysical
Research Letters,27, 2689– 2692.
OBE RHAN SLI , H., MCKEN ZIE, J., TOUMA RKI NE, M. &
WEI SSER T, H. 1984. A paleoclimatic and paleoceano-
graphic record of the Paleogene in the central South
Atlantic (Leg 73, Sites 522, 523, and 524). In: HS U
¨,
K. J. & LABR EQUE , J. L. ET AL . (eds) Initial
Reports DSDP, Leg 73, Washington DC, US Govern-
ment Printing Office, 737–747.
OBO H-IK UEN OBE , F. & JARA MIL LO, C. 2003. Palynolo-
gical patterns in uppermost Eocene to lower Oligocene
sedimentary rocks in the US Gulf Coast. In:
PRO THER O, D. R., IVA NY, L. & NESBI TT, E. A. (eds)
From Greenhouse to Icehouse. New York, Columbia
University Press, 269–282.
OER LEMA NS, J. 2004. Correcting the Cenozoic d
18
O
deep-sea temperature record for Antarctic ice
volume. Palaeogeography, Palaeoclimatology,
Palaeocecology,208, 195– 205.
OER LEMA NS, J. 2005. Antarctic ice volume and deep-sea
temperature during the last 50 Myr: A model study.
Annals of Glaciology,39, 13–19.
OLE INIK , A. E. & MAR INCO VIC H, L. J. 2003. Biotic
response to the Eocene-Oligocene transition: gastro-
pod assemblages in the high-latitude North Pacific.
In: PRO THE RO, D. R., IVAN Y, L. & NES BIT T, E. A.
H. K. COXALL & P. N. PEARSON382

(eds) From Greenhouse to Icehouse, New York,
Columbia University Press, 36– 56.
OLN EY, M. P., SCHER ER, R. P., BOHAT Y, S. M. &
HAR WOOD , D. M. 2005. Eocene– Oligocene paleoe-
cology and the diatom genus Kisseleviella Sheshu-
kova-Poretskaya from the Victoria Land Basin,
Antarctica. Marine Micropaleontology,58, 56– 72.
OPD YKE, B. N. & W ILK INS ON, B. H. 1989. Surface area
control of shallow cratonic to deep marine carbonate
accumulation. Paleoceanography,3, 685– 703.
PAL AZZE SI, L. & BARRE DA, V. 2007. Major vegetation
trends in the Tertiary of Patagonia (Argentina): A
qualitative paleoclimatic approach based on palynolo-
gical evidence. Flora: Morphology, Distribution,
Functional Ecology of Plants,202, 328–337.
PAG ANI, M. 2002. The alkenone-CO
2
proxy and
ancient atmospheric carbon dioxide. Philosophical
Transactions of the Royal Society of London,360,
609–632.
PAG ANI, M., L EMA RCH AND, D., S PIV ACK , A. & GAIL -
LA RDET , J. 2005a. A critical evaluation of the boron
isotope-pH proxy: The accuracy of ancient ocean pH
estimates. Geochimica et Cosmochimica Acta,69,
953–961.
PAG ANI, M., Z ACH OS, J. C., FREEM AN, K. H., T IPP LE,
B. & BO HATY , S. 2005b. Marked decline in atmos-
pheric carbon dioxide concentrations during the Paleo-
gene. Science,309, 600– 603.
PA
¨LI KE, H., LA SKA R, J. & SH ACKL ETO N, N. J. 2004.
Geologic constraints on the chaotic diffusion of the
solar system. Geology,32, 929– 932.
PA
¨LI KE, H., NO RRI S, R. N., HE RRLE , J. ET AL . 2006. The
heartbeat of the Oligocene climate system. Science,
314, 1894– 1898.
PAL MER, M. R. & E DMO ND, J. M. 1992. Controls over
the strontium isotope composition of river water.
Geochimica et Cosmochimica Acta,56, 2099– 2111.
PAN , A. D., JACO BS, B. F., DRAN SFIE LD, J. & BAKER ,
W. J. 2006. The fossil history of palms (Arecaceae)
in Africa and new records from the Late Oligocene
(28– 27 Ma) of north-western Ethiopia. Botanical
Journal of the Linnean Society,151, 69–81.
PAU L, H. A., ZAC HOS , J. C., FLO WER , B. P. & TRIP ATI ,
A. 2000. Orbitally induced climate and geochemical
variability across the Oligocene/Miocene boundary.
Paleoceanography,15, 471– 485.
PEA RSON , P. N. & CHA ISSO N, P. 1997. Late Paleocene to
middle Miocene planktonic foraminifer biostratigra-
phy of the Ceara Rise. In: SH ACK LETO N, N. J.,
CUR RY, W. B., RICHT ER, C. & BRALO WER , T. J.
(eds) Proceedings of the Ocean Drilling Program
Scientific Results,154, College Station, TX (Ocean
Drilling Program).
PEA RSON , P. N. & PAL MER, M. R. 1999. Middle Eocene
seawater pH and atmospheric carbon dioxide concen-
trations. Science,284, 1824– 1826.
PEA RSON , P. N. & PAL MER, M. R. 2000. Atmospheric
carbon dioxide concentrations over the past 60
million years. Nature,406, 695– 699.
PEA RSON , P. N., DIT CHF IELD , P. W., SIN GAN O, J.,
HAR COUR T-BROW N, K. G., NIC HOL AS, C. J.,
SHA CKLE TON , N. J. & HAL L, M. A. 2001. Warm tro-
pical sea surface temperatures in the Late Cretaceous
and Eocene epochs. Nature,413, 481– 487.
PEA RSON , P. N., OLS SON , R. K., HEM LEBE N, C.,
HUB ER, B. T. & BERGG REN , W. A. 2006. Atlas of
Eocene planktonic foraminifera. Cushman Foundation
of Foraminiferal Research, Special Publication: Lawr-
ence, Kansas, Allen Press.
PEA RSON , P. N., VAN DO NGE N, B. E., NI CHOL AS, C. J.,
PAN COST , R. D., SCH OUT EN, S., SINGA NO, J. &
WAD E, B. S. 2007. Stable warm tropical climate
through the Eocene epoch. Geology,35, 211–214.
PEG RAM, W. J. & TU REK IAN , K. K. 1999. The osmium
isotopic composition change of Cenozoic sea water
as inferred from a deep-sea core corrected for meteori-
tic contributions. Geochimica et Cosmochimica Acta,
63, 4053–4058.
PEK AR, S. & MILLE R, K. G. 1996. New Jersey Oligocene
‘icehouse’ sequences (ODP Leg 150X) correlated with
global d
18
O and Exxon eustatic records. Geology,24,
567–570.
PEK AR, S. F., CHRIS TIE -BLI CK, N., K OMI NZ, M. A. &
MIL LER, K. M. 2002. Calibration between
eustatic estimates from backstripping and oxygen
isotopic records for the Oligocene. Geology,30,
903–906.
PER CH-N IEL SEN , K. 1986. Calcareous nannofossil events
at the Eocene–Oligocene boundary. In: PO MER OL, C.
& PR EMOL I SILV A, I. (eds) Terminal Eocene Events,
Elsevier, 275–282.
PET IT, J. R., JOUZE L, J., RA YNAU D, D. ET A L. 1999.
Climate and atmospheric history of the past 420,000
years from the Vostok ice core, Antarctica. Nature,
399, 429– 436.
PET ERSO N, L. C. & BA CKM AN, J. 1990. Late Cenozoic
carbonate accumulation and the history of the carbon-
ate compensation depth in the western equatorial
Indian Ocean. In: BA CKMA N, J. & PE TER SON, L. C.
ET A L. (eds) Proceedings of the Ocean Drilling
Program, Scientific Results,115, College Station,
TX, Ocean Drilling Program, 467–489.
PEU CKER -EH REN BRIN K, B. & RA VIZ ZA, G. 2000. The
marine osmium isotope record. Terra Nova,12,
205–219.
PEU CKER -EH REN BRIN K, B., RA VIZ ZA, G. & HOFMA NN,
A. W. 1995. The marine
187
Os/
186
Os record of the past
80 millions years. Earth and Planetary Science
Letters,130, 155– 167.
POA G, C. W. 1995. Upper Eocene impactites of the U.S.
East Coast: Depositional origins, biostratigraphic
fram work, and correlation. Palaios,10, 16–43.
POA G, C. W., MAN KIN EN, E. & NO RRI S, R. 2003. Late
Eocene Impacts: geological record, correlation, and
paleoenvironmental consequences. In: PRO THE RO,
D. R., IV ANY, L. & N ESBI TT, E. A. (eds) From Green-
house to Icehouse: The Marine Eocene–Oligocene
Transition, New York, Columbia University Press,
495–510.
POL LARD , D. & DECON TO, R. M. 2003. Antarctic ice
and sediment flux in the Oligocene simulated by
a climate-ice sheet-sediment model. Palaeogeogra-
phy, Palaeoclimatology, Palaeoecology,198, 53– 67.
POL LARD , D. & DECON TO, R. M. 2005. Hysteresis in
Cenozoic Antarctic ice-sheet variations. Global and
Planetary Change,45, 9– 21.
POL LARD , D., DECON TO, R. M. & NYBLA DE, A. A.
2005. Sensitivity of Cenozoic Antarctic ice sheet
THE E–O TRANSITION 383

variations to geothermal heat flux. Global and Plane-
tary Change,49, 63– 74.
POM EROL , C. & PRE MOL I SILVA , I. 1986. Terminal
Eocene Events: Amsterdam, Elsevier, 412pp.
POO RE, R. Z. 1984. Middle Eocene through Quaternary
planktonic foraminifers from the southern Angola
Basin: Deep-sea Drilling Project Leg 73. In: HS U
¨,
K. J. & LABR EQUE , J. L. (eds) Initial Reports of
the Deep-Sea Drilling Project, Washington DC, US
Government Printing Office, 429–448.
PRE MOLI SI LVA , I. & JEN KINS , D. G. 1993. Decision on
the Eocene–Oligocene boundary stratotype. Episodes,
16, 379–381.
PRE NTIC E, M. L. & MA TTHE WS, R. K. 1988. Cenozoic
ice-volume history: development of a composite
oxygen isotope record. Geology,16, 963–966.
PRE NTIC E, M. L. & MA TTHE WS, R. K. 1991. Tertiary ice
sheet dynamics: The snow gun hypothesis. Journal of
Geophysical Research,96, 6811– 6827.
PRO THER O, D. R. 1994. The Eocene –Oligocene Tran-
sition: Paradise Lost. New York, Columbia University
Press, 291pp.
PRO THER O, D. R. & BE RGG REN , W. A. 1992. Eocene–
Oligocene Climatic and Biotic Evolution, Princeton
University Press, 568.
PRO THER O, D. & EM RY, R. J. 1996. The Terrestrial
Eocene–Oligocene Transition in North America,
Cambridge University Press, 688.
PRO THER O, D. R. & SW ISH ER, C. C. I. 1992. Magnetos-
tratigraphy and geochronology of the Terrestrial
Eocene–Oligocene transition in North America. In:
PRO THER O, D. R. & BE RGG REN, W. A. (eds)
Eocene–Oligocene Climatic and Biotic Evolution,
Princeton, Princeton University Press, 46–73.
PRO THER O, D. R., IV ANY , L. C. I. & NESB ITT , E. A.
2003. From Greenhouse to Icehouse: The Marine
Eocene–Oligocene Transition, Columbia University
Press, 560.
RAC K, F. R. 1993. A geologic perspective on the Miocene
evolution of the Antarctic Circumpolar Current
system. Tectonophysics,222, 397– 415.
RAG E, J. C. 1986. The amphibians and reptiles at the
Eocene–Oligocene boundary in western Europe: an
outline of the faunal alterations. In: PO MER OL, C. &
PRE MOLI SI LVA , I. (eds) Terminal Eocene Events,
Elsevier, 309–310.
RAM USSE N, D. T., BO WN, T. M. & SIMON S, E. L. 1992.
The Eocene–Oligocene transition in continental
Africa. In: PR OTHE RO, D. R. & BERG GREN , W. A.
(eds) Eocene– Oligocene Climatic and Biotic
Evolution, Princeton, Princeton University Press,
548–567.
RAV IZZA , G. & PEU CKE R-EH REN BRI NK, B. 2003.
The marine
187
Os/
188
Os record of the Eocene–
Oligocene transition: The interplay of weathering
and glaciation. Earth and Planetary Science Letters,
210, 151– 165.
RAY MO, M. E. 1991. Geochemical evidence supporting
T.C. Chamberlin’s theory of glaciation. Geology,19,
344–347.
RAY MO, M. E. & RUDDI MAN , W. F. 1992. Tectonic
forcing of late Cenozoic climate.Nature,359, 117– 122.
REA , D. K. & LYLE , M. W. 2005. Paleogene calcite
compensation depth in the eastern subtropical
Pacific: Answers and questions. Paleoceanography,
20, doi:10.1029/2004PA001064.
REA , D. K., LEIN EN, M. & J ANE CEK , T. R. 1985. Geo-
logic approach to the long-term history of atmospheric
circulation. Science,227, 721– 725.
REI LLY, T. J., M ILLE R, K. G. & FE IGE NSON , M. D. 2002.
Latest Eocene–earliest Miocene Sr isotopic reference
section, Site 522, eastern South Atlantic. Paleoceano-
graphy,17, doi:10.1029/2001PA000745.
RET ALLA CK, G. J. 2002. Carbon dioxide and climate over
the past 300 Myr. Philosophical Transactions of the
Royal Society of London. Series A,360, 659–673.
RIC HTER , F. M., ROW LEY , D. B. & DEPAO LO, D. J. 1992.
Sr isotope evolution of seawater: The role of tectonics.
Earth and Planetary Science Letters,109, 11–23.
RIE DEL, W. R. & S ANF ILIP PO, A. 1986. Radiolarian
events and the Eocene–Oligocene boundary. In:
POM EROL , C. & PRE MOL I SILV A, I. (eds) Terminal
Eocene Events, Elsevier, 253– 258.
ROB ERT, C., D IEST ER- HAA SS, L. & CH AML EY, H. 2002.
Late Eocene–Oligocene oceanographic development
at southern high latitudes, from terrigenous and bio-
genic particles: A comparison of Kerguelen Plateau
and Maud Rise, ODP Sites 744 and 689. Marine
Geology,191, 37– 54.
ROB INSO N, E. 2003. Upper Paleogene larger foramini-
feral succession on a tropical carbonate bank, Nicara-
gua Rise, Caribbean region. In: PR OTHE RO, D. R.,
IVA NY, L. & NE SBI TT, E. A. (eds) From Greenhouse
to Icehouse, New York, Columbia University Press,
294–302.
ROT H-NE BEL SIC K, A., UTE SCH ER, T., MO SBR UGG ER,
V., DI ESTE R-HAASS , L. & WAL THE R, H. 2004.
Changes in atmospheric CO
2
concentrations and
climate from the Late Eocene to Early Miocene:
Palaeobotanical reconstruction based on fossil floras
from Saxony, Germany. Palaeogeography, Palaeocli-
matology, Palaeoecology,205, 43– 67.
ROY ER, D. L. 2001. Stomatal density and stomatal
index as indicators of palaeoatmospheric CO
2
concen-
tration. Review of Palaeobotany and Palynology,114,
1–28.
ROY ER, D. L., WING, S. L., B EERL ING , D. J., JOLL EY, D.
W., KO CH, P. L., HI CKE Y, L. J. & BE RNER , R. A.
2001. Paleobotanical evidence for near present-day
levels of atmospheric CO
2
during part of the Tertiary.
Science,292, 2310– 2313.
ROY ER, D. L., BERNE R, R. A., MO NTA N
˜EZ , I. P., TABOR,
N. J. & BE ERLI NG, D. J. 2004. CO
2
as a primary driver
of Phanerozoic climate. GSA Today,14, 4– 10.
ROY ER, D. L., BE RNE R, R. A. & PAR K, J. 2007. Climate
sensitivity constrained by CO
2
concentrations over the
past 420 million years. Nature,446, 530–532.
RUD DIMA N, W. F., KU TZB ACH, J. E. & P REN TICE , I. C.
1997. Testing the climatic effects of orography and
CO
2
with general circulation and biome models. In:
RUD DIMA N, W. F. (ed.) Tectonic Uplift and Climate
Change, New York, Plenum 204–235.
SAG NOTT I, L., FL ORI NDO, F., V ERO SUB, K. L.,
WIL SON, G. S. & ROB ERT S, A. P. 1998. Environ-
mental magnetic record of Antarctic palaeoclimate
from Eocene–Oligocene glaciomarine sediments,
Victoria Land Basin. Geophysical Journal Inter-
national,134, 653– 662.
H. K. COXALL & P. N. PEARSON384

SAL AMY, K. A. & Z ACH OS, J. C. 1999. Latest Eocene –
Early Oligocene climate change and Southern Ocean
fertility: inferences from sediment accumulation and
stable isotope data. Palaeoceanography, Palaeoclima-
tology, Palaeoecology,145, 61– 67.
SAV AGE, D. E. & R USS ELL, D. E. 1983. Mammalian
Paleofaunas of the World: Reading, MA, Addison-
Wesley, 432pp.
SCH ELLE NBE RG, S. A. 1998. Ecological and evolutionary
response of deep-sea ostracodes to the Eocene –Oligo-
cene transition: High-resolution data from ODP Site
744, Kerguelen Plateau. GSA Annual Meeting
Abstracts with Programs, A-286.
SCH ER, H. D. & MA RTI N, E. E. 2004. Circulation in the
Southern Ocean during the Paleogene inferred from
neodymium isotopes. Earth and Planetary Science
Letters,228, 391– 405.
SCH ER, H. D. & MA RTI N, E. E. 2006. Timing and cli-
matic consequences of the opening of Drake
Passage. Science,312, 428– 430.
SCH OLL, D. W., S TEV ENS ON, A. J., NO BLE , M. A. &
REA , D. K. 2003. The Meiji Drift Body and Late
Paleogene– Neogene Paleoceanography of the North
Pacific–Bering Sea region. In: PR OTH ERO, D. R.,
IVA NY, L. & NESBI TT, E. A. (eds) From Greenhouse
to Icehouse, New York, Columbia University Press,
119–153.
SCH OUTE N, S., HO PMA NS, E. C., SC HEF U, E. &
SIN NING HE DAM STE
´, J. S. 2002. Distributional vari-
ations in marine crenarchaeotal membrane lipids: A
new tool for reconstructing ancient sea water
temperatures? Earth and Planetary Science Letters,
204, 265– 274.
SCH UMAC HER , S. & LAZ ARU S, D. 2004. Regional differ-
ences in pelagic productivity in the late Eocene to
early Oligocene-a comparison of southern high lati-
tudes and lower latitudes. In: KI ESS LING , W. &
LAZ ARUS , D. (eds) Mesozoic– Cenozoic Bioevents,
Palaeogeography, Palaeoclimatology, Palaeoecol-
ogy, 243– 263.
SE
´RA NNE, M. 1999. Early Oligocene stratigraphic turn-
over on the west Africa continental margin: a signature
of the Tertiary greenhouse-to-icehouse transition?
Terra Nova,11, 135–140.
SEP KOSK I, J. J. J. 1986. Phanerozoic overview of mass
extinction. In: RA UP, D. M. & JA BLO NSK I, D. (eds)
Patterns and Processes in the History of Life, Berlin,
Springer Verlag, 277–295.
SHA CKLE TON , N. J. & KENN ETT , P. 1975. Paleotempera-
ture history of the Cenozoic and the initiation of
Antarctic glaciation: oxygen and carbon isotope analy-
sis in DSDP Sites 277, 279 and 281. In: KENNE TT,
J. P. & HO UTZ, R. E. E T AL. (eds) Initial Reports of
the Deep-Sea Drilling Project 29, Washington DC,
US Government Printing Office, 881–884.
SHA CKLE TON , N. J. & OPDYKE, N. D. 1973. Oxygen
isotope and paleomagnetic stratigraphy of equatorial
Pacific core V28–238: Oxygen isotope temperatures
and ice volumes on a 10
5
and 10
6
year scale. Quatern-
ary Research,3, 39–55.
SHI PBOA RD SCI ENTI FIC PA RTY ,ET AL. 2001. Leg 189
summary. In: EXO N, N. F., KE NNE TT, J. P.,
MAL ONE, M. J. (eds) Proceedingsof the Ocean Drilling
Program Initial Reports, 189, College Station, TX,
Ocean Drilling Program, [Online]. Available from:
http://www- odp.tamu.edu/publica tions /189_I R/
chap_01/chap_01. htm
SIG MAN, D. & B OYLE , E. 2000. Glacial/Interglacial vari-
ations in atmospheric carbon dioxide. Nature,407,
859–869.
STI CKLE Y, C. E., BR INK HUI S, H., SCH ELL ENBE RG, S. A.
ET A L. 2004. Timing and nature of the deepening of the
Tasmanian Gateway. Paleoceanography,19, doi:10.
1029/2004PA001022.
SLU IJS, A., B RIN KHU IS, H., ST ICK LEY , C. E.,
WAR NAAR , J. & WIL LIA MS, G. L. 2003. Dinoflagel-
late cysts from the Eocene –Oligocene transition in the
Southern Ocean; results from ODP Leg 189. In: EX ON,
N. F., KE NNET T, J. P. & MA LONE , M. J. (eds) Pro-
ceedings of the Ocean Drilling Program Scientific
Results ODP Leg 189, College Station, TX, Ocean
Drilling Program, 1–42 [Online]. Available from:
htt p://ww w://ww w.od p.tamu.edu/p ubli cati ons/
189_SR/VOLUME/CHAPTERS/104. PDF.
SQU IRES , R. L. 2003. Turnovers in marine gastropod
assemblages in the high-latitude North Pacific. In:
PRO THER O, D. R., IV ANY , L. & NES BITT , E. A.
(eds) From Greenhouse to Icehouse, New York,
Columbia University Press, 14–35.
STE HLIN , H. G. 1909. Remarques sur les faunules de
Mammife
`res des couches Eoce
`nes et Oligoce
`nes du
Bassin de Paris. Bulletin de la Socie
´te
´Ge
´ologique de
France,29, 488– 520.
STO TT, L. D., KENNE TT, J. P., S HAC KLET ON, N. J. &
COR FIEL D, R. M. 1990. The evolution of Antarctic
surface waters during the Paleogene: Inferences from
the stable isotopic composition of planktonic foramini-
fers, ODP Leg 113. In: BA RKE R, P. F., KEN NET T, J. P.
ET A L. (eds) Proceedings of the Ocean Drilling
Program, Scientific Results, College Station, TX,
Ocean Drilling Program, 849–863.
STR O
¨MB ERG, C. A. E. 2005. Decoupled taxonomic radi-
ation and ecological expansion of open-habitat
grasses in the Cenozoic of North America. Geology,
102, 11980– 11984.
STR O
¨MB ERG, C. A. E. 2006. Evolution of hypsodonty in
equids: testing a hypothesis of adaptation. Palaeo-
biology,32, 236– 258.
STU CKY, R. K. 1992. Mammalian faunas in North
America of Bridgerian to early Arikareean ‘ages’
(Eocene and Oligocene). In: PR OTHE RO, D. R. &
BER GGRE N, W. A. (eds) Eocene –Oligocene Climatic
and Biotic Evolution, 464– 493.
SUT O, I. 2006. The explosive diversification of the diatom
genus Chaetoceros across the Eocene–Oligocene and
Oligocene– Miocene boundaries in the Norwegian
Sea. Marine Geology,58, 259– 269.
TAK ATA, H., N OMU RA, R. & SETO , K. 2007. Review of
paleoceanographic transition during the Oligocene in
the eastern equatorial Pacific Ocean. Fossils,81,
15–14.
THO MAS, E. 1992. Middle Eocene –late Oligocene
bathyal benthic foraminifera (Weddell Sea):
faunal changes and implications for ocean circulation.
In: PRO THE RO, D. R. & BERGG REN , W. A.
(eds) Eocene–Oligocene Climatic and Biotic Evol-
ution, Princeton, NY, Princeton University Press,
245–271.
THE E–O TRANSITION 385

THO MAS, E. & G OOD AY, A. J. 1996. Cenozoic deep-sea
benthic foraminifers: tracers for changes in oceanic
productivity? Geology,24, 355– 358.
THO RN, V. C. & DECO NTO , R. 2006. Antarctic climate at
the Eocene–Oligocene boundary – Climate model
sensitivity to high latitude vegetation type and com-
parisons with the palaeobotanical record. Palaeogeo-
graphy, Palaeoclimatology, Palaeoecology,231.
THU NELL , R. & COR LIS S, B. H. 1986. Late Eocene –early
Oligocene carbonate sedimentation in the deep-sea. In:
POM EROL , C. & PRE MOLI SI LVA , I. (eds) Terminal
Eocene Events, Elsevier, 363–380.
TOG GWEI LER , J. R. & BJOR NSS ON, H. 2000. Drake
Passage and paleoclimate. Journal of Quaternary
Science,15, 319– 328.
TRI PATI , A., BAC KMA N, J., ELD ERF IEL D, H. &
FER RETT I, P. 2005. Eocene bipolar glaciation associ-
ated with global carbon cycle changes. Nature,436,
341–346.
TUO , S.-T., LIU, Z.-F., Z HAO , Q.-H. & CHE NG, X.-R.
2006. Earliest Oligocene glacial maximum: records
from ODP Site 1265, South Atlantic. Diqiu Kexue –
Zhongguo Dizhi Daxue Xuebao/Earth Science –
Journal of China University of Geosciences,31,
151–158.
VAI L, P. R. & HAR DEN BOL, J. 1979. Sea level changes
during the Tertiary. Oceanus,22, 171– 179.
VAN AN DEL, T. H. 1975. Mesozoic/Cenozoic calcite
compensation depth and the global distribution of cal-
careous sediments. Earth and Planetary Science
Letters,26, 187– 194.
VA N DEN BOL D, W. A. 1986. Ostracodes at the Eocene –
Oligocene Boundary in the Aquitaine basin. Stratigra-
phy, phylogeny, palaeoenviornments. In: PO MERO L,
C. & PR EMOL I SILV A, I. (eds) Terminal Eocene
Events, Elsevier, 259– 264.
VAN DON GEN, B. E., T ALB OT, H. M., SCHOU TEN , S.,
PEA RSON , P. N. & PAN COST , R. D. 2006. Well pre-
served Palaeogene and Cretaceous biomarkers from
the Kilwa area, Tanzania. Organic Geochemistry,37,
539–557.
VAN MOU RIK, C. A. & B RIN KHUI S, H. 2005. The Mas-
signano Eocene–Oligocene golden spike section
revisited. Stratigraphy,2, 13– 29.
VAN DEN BER G, M. D. & JA RRA RD, R. D. 2004. Ceno-
zoic mass accumulation rates in the equatorial
Pacific based on high-resolution mineralogy of
Ocean Drilling Program Leg 199. Paleoceanography,
19, doi:10.1029/2003PA000928.
VAN DENB ERG HE, N., BRINK HUI S, H. & STE URB AUT , E.
2003. The Eocene–Oligocene boundary in the North
Sea Area: a sequence stratigraphic approach. In:
PRO THER O, D. R., IV ANY , L. C. & NESB ITT , E. A.
(eds) From Greenhouse to Icehouse, New York,
Columbia University Press, 419–438.
VAR OL, O. 1998. Paleogene. In: BO WN, P. R. (ed). Cal-
careous Nannofossil Biostratigraphy, Dordrecht,
Kluwer Academic Publishing, 201–224.
VER GNAU D-GRAZZ INI , C. & OBE RHA NSLI , H. 1986.
Isotopic events at the Eocene Oligocene boundary –
A review. In: PO MERO L, C. & PR EMO LI SIL VA, I.
(eds) Terminal Eocene Events, Elsevier, 311–330.
VIA , R. K. & THOM AS, D. J. 2006. Evolution of Atlantic
thermohaline circulation: Early Oligocene onset of
deep-water production in the North Atlantic.
Geology,34, 441– 444.
VOL K, T. 1987. Feedbacks between weathering and
atmospheric CO
2
over the last 100 million years.
American Journal of Science,287, 763– 779.
VON HOF, H. B., S MIT , J., BRI NKHU IS, H., MONT ANAR I,
A. & NE DERB RAG T, A. J. 2000. Global cooling accel-
erated by early late Eocene impacts? Geology,28,
267–293.
WADE, B. S. & PA
¨LI KE, H. 2004. Oligocene climate
dynamics. Paleoceanography,19, PA4019, doi:
10.1029/2004PA001042.
WHI TEHE AD, J. M. 2005. Cenozoic history of Antarctic
benthic diatoms. Alcheringa,29, 151–169.
WIL SON, G. S., R OBE RTS , A. P., VER OSUB , K. L.,
FLO RIND O, F. & SA GNO TTI, L. 1998. Magnetobios-
tratigraphic chronology of the Eocene–Oligocene
transition in the CIROS-1 core, Victoria Land
margin, Antarctica: Implications for Antarctic glacial
history. Bulletin of the Geological Society of
America,110, 35– 47.
WIL SON, P. A. & N ORR IS, R. D. 2001. Warm tropical
ocean surface and global anoxia during the mid-
Cretaceous period. Nature,412, 425– 429.
WIS E, S. W. JR., BR EZA , J. R., HAR WOOD , D. M. & WEI ,
W. 1991. Paleogene glacial history of Antarctica. In:
MUE LLER , D. W., MCKEN ZIE , J. A. & WEIS SER T,
H. (eds) Controversies in Modern Geology; Evolution
of Geological Theories in Sedimentology, Earth
History and Tectonics, London, Academic Press,
133–171.
WOL FE, J. A. 1978. A paleobotanical interpretation of
Tertiary climates in the northern hemisphere. Ameri-
can Scientist,66, 694– 703.
WOL FE, J. A. 1992. Climatic, floristic, and vegetational
changes near the Eocene–Oligocene boundary in
North America. In: PR OTHE RO, D. R. & BERG GREN ,
W. A. (eds) Eocene–Oligocene Climatic and Biotic
Evolution, Princeton, NJ, Princeton University Press,
421–436.
WOL FE, J. A. 1994. Tertiary climatic changes at
middle latitudes of western North America. Palaeo-
geography, Palaeoclimatology, Palaeoecology,108,
195–205.
WRI GHT, J., S EYM OUR, R. S. & S HAW , H. 1984. REE
and Nd isotopes in conodont apatite: Variations with
geological age and depositional environment. In:
CLA RK, D. L. (ed.) Conodont Biofacies and Provinci-
alism: Geological Society of America Special Paper
196, 325– 340.
WRI GHT, A. & M ILL ER, K. G. 1993. Southern ocean
influences on late Eocene to Miocene deepwater circu-
lation. The Antarctic paleoenvironment: A perspective
on global change Antarctic research series, 1–25.
YAN CEY, T. E., E LSI K, W. C. & SA NCY, R. H. 2003. The
Palynological record of the Late Eocene climate
change, Northwest Gulf of Mexico. In: PR OTH ERO,
D. R., IV ANY, L. & N ESBI TT, E. A. (eds) From Green-
house to Icehouse, New York, Columbia University
Press, 252–268.
ZAC HOS, J. C. & K UMP , L. R. 2005. Carbon cycle feed-
backs and the initiation of Antarctic glaciation in the
earliest Oligocene. Global and Planetary Change,
47, 51– 66.
H. K. COXALL & P. N. PEARSON386

ZAC HOS, J. C., B REZ A, J. R. & WI SE, S. W. 1992. Early
Oligocene ice sheet expansion on Antarctica. Geology,
569–573, 839–854.
ZAC HOS, J. C., Q UIN N, T. M. & SA LAMY , K. A.
1996. High-resolution (10
4
years) deep-sea forami-
niferal stable isotope records of the Eocene –Oligo-
cene climate transition. Paleoceanography,11,
251–266.
ZAC HOS, J. C., FL OWE R, B. P. & PAU L, H. 1997.
Orbitally paced climate oscillations across the
Oligocene/Miocene boundary. Nature,388,
567–570.
ZAC HOS, J. C., O PDY KE, B. N., QUINN , T. M., JON ES, C.
E. & HA LLID AY, A. N. 1999. Early Cenozoic
glaciation, Antarctic weathering, and seawater
87
Sr/
86
Sr; is there a link? Chemical Geology,161,
165–180.
ZAC HOS, J. C., P AGA NI, M., SLOAN , L. C., THO MAS , E.
& BI LLUP S, K. 2001. Trends, rhythms, and aberra-
tions in global climate 65 Ma to present. Science,
292, 686– 693.
ZAN AZZI , A., KOH N, M. J., MA CFADD EN, B. J. &
TER RY, D. O. 2007. Large temperature drop across
the Eocene–Oligocene transition in central North
America. Nature,445, 639–642.
ZEE BE, R. E. & WE STB ROE K, P. 2003. A simple model
for the CaCO
3
saturation state of the ocean: The
‘Strangelove’, the ‘Neritan’, and the ‘Cretan’ ocean.
Geochemistry, Geophysics, Geosystems,4, doi:
10.1029/2003GC000538.
THE E–O TRANSITION 387