Were transgressive black shales a negative feedback modulating glacioeustasy in the Early Palaeozoic Icehouse?

Abstract

The Early Palaeozoic Icehouse (Late Ordovician-Early Silurian, c. 455-425 Ma) was a remarkable event in the Earth's climatic history, marked by extensive glaciations occurring at a time of elevated atmospheric CO2. The oceanography of the Early Palaeozoic Icehouse was markedly different from that of modern oceans, with frequent episodes of oceanic anoxia and high concentrations of CO2 which may have acidified the oceans and restricted carbonate burial. Thus, the marine organic carbon reservoir may have more strongly influenced long-term changes in atmospheric CO2 than at present. We suggest that deposition of black shales represented a major sink for atmospheric carbon. Sequence stratigraphy reveals that widespread black shale deposition occurred in transgressions, whereas regressions are characterized by deposition of bioturbated facies, allowing changes in lithofacies and deep-water redox conditions to be related to the Early Palaeozoic carbon cycle. Assuming increased temperature is a function of increased atmospheric CO2, and that glacioeustatic sea-level can serve as a proxy for temperature due to changing ice volume, we infer that the deposition of transgressive black shales may have acted as a negative feedback mechanism, drawing down CO2 and preventing the onset of runaway greenhouse conditions.

Full text

Were transgressive black shales a negative feedback modulating
glacioeustasy in the Early Palaeozoic Icehouse?
A. A. PAGE1,2, J. A. ZALASIEWICZ1, M. WILLIAMS1& L. E. POPOV3
1
Department of Geology, University of Leicester, Leicester, LE1 7RH, UK
(e-mail: aap30@esc.cam.ac.uk)
2
Department of Earth Sciences, University of Cambridge, Downing Street,
Cambridge, CB2 3EQ, UK
3
National Museum of Wales, Department of Geology, Cathays Park,
Cardiff CF10 3NP, UK
Abstract: The Early Palaeozoic Icehouse (Late Ordovician–Early Silurian, c. 455– 425 Ma) was
a remarkable event in the Earth’s climatic history, marked by extensive glaciations occurring at a
time of elevated atmospheric CO
2
. The oceanography of the Early Palaeozoic Icehouse was
markedly different from that of modern oceans, with frequent episodes of oceanic anoxia and
high concentrations of CO
2
which may have acidified the oceans and restricted carbonate
burial. Thus, the marine organic carbon reservoir may have more strongly influenced long-term
changes in atmospheric CO
2
than at present. We suggest that deposition of black shales rep-
resented a major sink for atmospheric carbon. Sequence stratigraphy reveals that widespread
black shale deposition occurred in transgressions, whereas regressions are characterized by depo-
sition of bioturbated facies, allowing changes in lithofacies and deep-water redox conditions to be
related to the Early Palaeozoic carbon cycle. Assuming increased temperature is a function of
increased atmospheric CO
2
, and that glacioeustatic sea-level can serve as a proxy for temperature
due to changing ice volume, we infer that the deposition of transgressive black shales may have
acted as a negative feedback mechanism, drawing down CO
2
and preventing the onset of runaway
greenhouse conditions.
The Early Palaeozoic represents an important inter-
val in Earth biosphere evolution. It post-dated the
origin of large metazoans and complex, tiered
marine food webs (Butterfield 1997), and is suc-
ceeded by the radiation of land plants (Berner
1998; Gensel & Edwards 2001). It therefore
represents an intermediate state between the
oxygen-poor Proterozoic palaeoenvironment and
the well-oxygenated world of the Late Palaeozoic
and post-Palaeozoic (Berner 2003; Catling &
Claire 2005). This interval marks a non-actualistic
solution to the Earth’s carbon budget. Though
generally considered an interval of long-lived,
stable greenhouse conditions (e.g. Gibbs et al.
2000; Montan
˜ez 2002; Church & Coe 2003,
fig. 5.4), major glaciations nonetheless occurred in
the late Ordovician and early Silurian. These glacia-
tions occurred at elevated atmospheric CO
2
(Royer
2006) and transitions between oxic and anoxic
marine conditions were frequent (Figs 1d & 2). In
a time before the evolution of a complex land
biota, most of the organic carbon reservoir must
have existed in oceans, where it was buried as
black shale (Fig. 1). Despite recent advances in
general circulation models (GCMs) and the appli-
cation of climatically sensitive stable isotopes to
infer palaeoenvironmental change, Early Palaeo-
zoic climate remains somewhat enigmatic, no
doubt in part due to its lack of analogue in the
modern world.
Instead, the Early Palaeozoic needs to be under-
stood in its own terms. Much as Charles Lapworth,
Adam Sedgwick and Roderick Murchison carefully
unpicked the undifferentiated ‘greywacke’ succes-
sions mapped by William Smith and Charles Lyell
in the 19th century, and established the stratigraphic
divisions of the Lower Palaeozoic (Rudwick 1985;
Secord 1986; Oldroyd 1990), the 21st century sees
the need for Early Palaeozoic workers to return to
its stratigraphy and establish how global lithostrati-
graphic patterns of continental weathering and car-
bonate and black shale burial relate to its
palaeoclimate, thereby determining the large-scale
controls on the carbon cycle at this time.
The Early Palaeozoic carbon cycle and
climate
Understanding chemical oceanography and carbon
cycling in the Early Palaeozoic is difficult. The
precise magnitude of atmospheric CO
2
at this time
From:WILLIAMS, M., HAYWOOD, A. M., GREGORY,F.J.&SCHMIDT, 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, 123 –156.
1747-602X/07/$15.00 #The Micropalaeontological Society 2007.

is uncertain, and the relation between CO
2
regu-
lation and Early Palaeozoic climate is not fully
resolved (see discussions in Ridgwell 2005; Royer
2006). However, available proxy data agree well
with Berner and Kothavala’s (2001) GEOCARB
III model of atmospheric CO
2
levels over Phanero-
zoic time (Crowley & Berner 2001; Royer et al.
2004; Royer 2006), providing support for extremely
elevated CO
2
levels in this interval (Fig. 1a). This
provides support for key assumptions of the
GEOCARB model and its descendants. Among
these assumptions is that long-term drawdown of
atmospheric CO
2
into the oceans was a conse-
quence of (a) continental silicate weathering and
burial in carbonates, and (b) photosynthesis and
burial of organic carbon (Berner 1991, 1994,
2006; Berner & Kothavala 2001). CO
2
regulation
must have been reflected in the specific pattern of
organic and inorganic carbon burial in the Early
Palaeozoic (Fig. 1), which differs notably from
that of the Neoproterozoic (Rothman et al. 2003)
and the rest of the Palaeozoic (Berner 2003).
In the Early Palaeozoic, carbonate burial was
restricted to the continental shelves (Walker et al.
2002), whilst organic carbon was predominantly
buried in deep-water anoxic environments
(Fig. 1d). The advent of biomineralization in the
Cambrian explosion facilitated carbonate burial
relative to the Neoproterozoic (Rothman et al.
2003; Ridgwell 2005). However, the Early Palaeo-
zoic may have lacked a well-developed marine car-
bonate buffer, which is highly sensitive to increased
atmospheric CO
2
(Barker et al. 2003). The radiation
of the calcifying plankton in the Triassic profoundly
affected the pattern of marine carbonate deposition
(Martin 1995), and a modern ocean analogue for
carbonate burial may not be applicable before this
(cf. Ridgwell 2005). Likewise, the advent of
Fig. 1. Graphs illustrating changes in the nature of the carbon cycle through Phanerozoic time, showing the Early
Palaeozoic dominated by organic-carbon burial in deep-marine anoxic waters under conditions of elevated atmospheric
CO
2
.(a) Temporal changes in atmospheric CO
2
relative to pre-industrial levels and partial pressure of atmospheric
O
2
(after Berner 2001; Berner & Kothavala 2001). (b) Reduced carbonate deposition during the EPI as recorded in
the relative proportion of low-latitude (,308N/S) shelf area occupied by carbonate sediments with time (data from
Walker et al. 2002). (c) Temporal changes in organic carbon burial flux, showing a notable high in the EPI when organic
carbon was predominantly buried in black shales (after Berner 2003). (d) Ratio of the accumulation rate of organic
carbon and pyrite–sulphur (C/S) in sediments versus time (after Berner 2003): low C/S values reflect deposition
of organic carbon in euxinic basins, high values correspond to burial in terrestrial freshwater swamps, and intermediate
values are found in normal marine sediments (Berner & Raiswell 1983).
A. A. PAGE ET AL.124

digestion, bioturbation and a macroplankton faecal
express in the Cambrian explosion (Butterfield
1997) no doubt significantly affected the cycling
of the organic carbon reservoir, which lacked the
sustained, large-scale fluctuations witnessed in the
Neoproterozoic (Hayes et al. 1999; Rothman et al.
2003). However, prior to the greening of the conti-
nents in the Late Palaeozoic (Berner 1998; Gensel
& Edwards 2001), the marine organic carbon
reservoir dominated the burial of organic carbon
(Berner 2003), and the frequent intervals of
marine anoxia that typify the Early Palaeozoic are
probably a consequence of this (Figs 1d & 3).
Constraining CO
2
drawdown in this interval
requires good estimates of the burial fluxes of car-
bonates and organic carbon (Fig. 1b, c). There
are two ways of achieving such estimates. The
first approach takes the known volume of carbon
held in preserved strata and applies a correction to
account for the progressive volume loss due
to erosion, subduction and metamorphism (e.g.
Berner & Canfield 1989; Walker et al. 2002).
The second approach applies GCMs and/or geo-
chemically appropriate mass-balance models to
proxy and/or mass flux curves (e.g. Berner 2003;
Locklair & Lermann 2005; Ridgwell 2005).
Neither of these methods is without problems: the
former depending heavily on the dataset and the
latter depending on the assumptions of the model.
Though sophisticated approaches such as GCMs
or multifactor box modelling allow palaeoclimatic
hypotheses to be quantitatively established and/or
tested, they may be extremely sensitive to certain
parameters and differing algorithms can produce
different results based on similar datasets (see
Haywood et al. 2005). Moreover, GCMs are
highly dependent on changes in palaeogeography,
ocean bathymetry, pCO
2
, insolation and albedo.
So, unless these factors are well constrained, their
results should be considered conservatively before
universally accepting their applicability in non-
actualistic environments (cf. Ridgwell 2005).
The Early Palaeozoic climate includes the see-
mingly paradoxical occurrence of extensive glacia-
tions (the Early Palaeozoic Icehouse or EPI as
defined below) at elevated atmospheric CO
2
(Royer 2006). Given the long-recognized coupling
of CO
2
and temperature (Arrhenius 1896;
Chamberlin 1899) and more recent affirmations
of the sensitivity of temperature to CO
2
(e.g.
Shackleton 2000; Zachos et al. 2001; Kump 2002;
Siegenthaler et al. 2005), a link between CO
2
and
temperature in the Early Palaeozoic seems reason-
able. Decreased cosmic ray flux also may have
contributed to globally cooler temperatures during
the EPI (Veizer et al. 2000; Shaviv 2002; Shaviv
& Veizer 2003), but this was insufficient to
induce glaciation alone (Royer 2006).
Most models suggest that atmospheric CO
2
was
the key control on temperature and ice formation in
the Ordovician and Silurian, predicting a pCO
2
-ice
threshold around 3000 ppm (Kump et al. 1999;
Hermann et al. 2003, 2004a,b; Royer 2006).
These values are significantly lower than the
GEOCARB III or GEOCARBSULF estimate for
Hirnantian CO
2
levels at c. 4000 ppm (Berner &
Kothavala 2001; Berner 2006). The GEOCARB/
GEOCARBSULF estimates are consistent with
estimates from goethite (Yapp & Poths 1992) and
significantly lower than the single palaeosol-based
estimate of CO
2
for the Ashgill at c. 5600 ppm
(Royer 2006). However, these models operate on
longer timescales than the duration of the short-
lived Hirnantian glacial maximum (cf. Sutcliffe
et al. 2000), so the discrepancy between the esti-
mated CO
2
-ice threshold and estimates of atmos-
pheric CO
2
may not be inconsistent (cf. Royer
2006). That is, glacial events could have been too
rapid to be captured by either these models or the
sparse proxy record. Individual glaciations in the
EPI may have been short-lived events related to
rapid CO
2
drawdown and cooling (e.g. Kump
et al. 1999).
A stratigraphic approach to Early
Palaeozoic glaciations
Lithostratigraphic correlation may establish a link
between deposition of CO
2
sinks and glaciations
during the Early Palaeozoic. CO
2
may be drawn
down from the atmosphere and sequestered in
rocks by (a) photosynthesis and burial of organic
carbon, or (b) continental silicate weathering and
carbonate deposition. In glacial intervals, changes
in ice-volume allow changes in glacioeustatic sea-
level to serve as a proxy for atmospheric CO
2
(assuming that ice volume was a decreasing func-
tion of temperature and that temperature was an
increasing function of atmospheric CO
2
). The stra-
tigraphic occurrence of carbonates and argillaceous
sediments has been well documented in the identifi-
cation of Primo/Secundo or Humid/Arid episodes
(e.g. Jeppsson 1990, 1997; Aldridge et al. 1993;
Jeppsson et al. 1995; Bickert et al. 1997; Cramer
& Saltzman 2007; see also discussion). We adopt
a complementary approach by comparing the strati-
graphic distributions of black shale with glacioeu-
static sea-level curves, isotopic data and evidence
of glaciations.
This approach depends on the selection of high
quality datasets with well-resolved stratigraphies
and accurate correlations. These are discussed in
the Appendix. We have been cautious in assigning
evidence of ice formation to the glacial maxima,
as discussed below.
BLACK SHALES IN THE EARLY PALAEOZOIC ICEHOUSE 125

A. A. PAGE ET AL.126

The Early Palaeozoic Icehouse
The Early Palaeozoic Icehouse (EPI) was an
approximately 30-million-year interval comprising
seven currently recognized glacial maxima
(Table 1; Fig. 2). We propose that the EPI began
with the Guttenberg Limestone carbon isotope
excursion (GICE) in the Caradoc (Ordovician),
and ended with Ireviken event deglacial transgres-
sion of the earliest Wenlock (Silurian). The EPI
reached its greatest extent in the short-lived Hirnan-
tian event identified by Brenchley et al. (1994), but
there is a good evidence for extensive ice formation
and significant glacioeustatic change throughout the
EPI (Tables 1 & 2; Fig. 2). Several other authors
have argued for an extended period of glaciation
(e.g. Frakes et al. 1992; Eyles 1993; Evans 2003;
Ghienne 2003; Kaljo et al. 2003; Nielsen 2003a;
Saltzman & Young 2005).
The sedimentary record of glaciations in the
EPI is predominantly held on Gondwana, with
Ordovician deposits generally found in Africa,
and Silurian deposits in South America (Table 2;
Eyles 1993; Dı
´az-Martı
´nez & Grahn 2007). This
continental-scale diachronism may reflect the
movement of Gondwana across the South Pole
(cf. Fortey & Cocks 2003). The Tamadjert Fm of
Saharan Africa displays a 200 m sequence of
glacial deposits. These include extensive tillites
and diamictites and abundant evidence of glacial
erosion stretching from the Caradoc to the late
Llandovery or even the earliest Wenlock (Beuf
et al. 1971; Biju-Dival et al. 1981 and references
therein). Three periods of continent-wide diamictite
deposition are recognized in the early Silurian of
South America (Caputo 1998; Dı
´az-Martı
´nez
2007). After the latest Llandovery, the South
American record of glaciation becomes ambiguous.
Glaciogenic diamictites in the San Gaba
´n–Canca-
n
˜iri–Zapla and Nhamunda
´Fms (Table 2) are over-
lain by strata yielding early Wenlock conodonts
and late Telychian– early Wenlock chitinozoans
respectively (Dı
´az-Martı
´nez 2007; Grahn in
Cramer & Saltzman 2007). The most recent works
on these glacial deposits consider them as having
an entirely Llandovery age (Dı
´az-Martı
´nez 2007;
Dı
´az-Martı
´nez & Grahn 2007). Though the Kirusil-
las Fm of Bolivia contains diamictites of early
Wenlock age (Merino 1991; Dı
´az-Martı
´nez 2007),
these lack glacially abraded clasts and are considered
to be sediment gravity flows (Dı
´az-Martı
´nez 2007;
Dı
´az-Martı
´nez & Grahn 2007).
The coupled, rapid variation in the isotopic and
glacioeustatic records (Fig. 2) that continues
throughout the EPI clearly marks a genetic change
in the Earth system. The strong co-variation in
these records may arise from the interval being a
prolonged icehouse event (Kaljo et al. 2003;
Nielsen 2003b). The good correspondence of
third-order variation in eustatic sea-level curves
(Ross & Ross 1996; Nielsen 2003a,b), along with
evidence of ice advance and retreat (Table 2)
argues for ice-volume controlling sea-level during
the EPI. Likewise, there is strong co-variation in
d
13
C and d
18
O data throughout the EPI (cf. Azmy
et al. 1998; Shields et al. 2003). During the EPI,
we interpret positive d
13
C excursions as being due
to increased weathering of shallow carbonates
exposed in regressions (cf. Kump et al. 1999;
Melchin & Holmden 2006; see also the Appendix).
Whilst in the EPI, positive d
18
O excursions
are interpreted as due to cooling rather than
increased salinity alone (Azmy et al. 1998). There-
fore, coupled, positive d
13
C and d
18
O excursions
may indicate glaciations if consistent with
other evidence.
The synchronous onset of significant, coupled
isotopic and eustatic fluctuations at the onset of
the Katian in the Ordovician, mark the beginning
of the EPI (Fig. 2; Patzkowsky et al. 1997; Kaljo
et al. 2003; Nielsen 2003a,b). This corresponds
to the mid-Caradoc clingani graptolite Zone
(Cooper & Sadler 2004; Goldman et al. 2005).
There is, though, also evidence of glacial erosion
and a regression in the earliest Caradoc
(Hamoumi 1999; Nielsen 2003a,b). The termin-
ation of the EPI is marked by the decoupling of iso-
topic and glacioeustatic variation in the earliest
Wenlock (Fig. 2). In the latest Telychian, there is
good evidence of ice (Table 2). Prior to the early
Fig. 2. (Opposite) The relationship between extent of marine anoxia and sea-level during the EPI based on
the chronostratigraphy of Cooper & Sadler (2004) and Melchin et al. (2004), but placing the Llandovery– Wenlock
Boundary at Ireviken datum 2 after the recommendations of Loydell et al. (2003) and Calner et al. (2004).
Transgressive anoxia is apparent by comparing (a) summary sea-level curves to (b) the oxic/anoxic stratigraphy (dark
bands, anoxic; white, oxic). Sea-level curves drawn after Ross & Ross (1996) and Nielsen (2003a); the oxic/
anoxic stratigraphy is compiled from Figure 3. The Ross & Ross (1996) curve has been modified in accordance with
its recorrelation in Loydell (1998). (c–f) Stable carbon-isotope curves: (c, d) redrawn from Kaljo et al. (2003);
(e) redrawn from Kaljo et al. (2004); (f) redrawn from Ainsaar et al. (1999). (g) Glacial maxima recognized in
Table 1: GICE, Guttenberg regression; ERR, early Rakvere regression; EAR, early Ashgill regression; HICE,
Hirnantian glaciation; EAGL, early Aeronian glaciation; SEDG, sedgwickii Zone glaciation; LTG, late
Telychian glaciation.
BLACK SHALES IN THE EARLY PALAEOZOIC ICEHOUSE 127

A. A. PAGE ET AL.128

Wenlock Ireviken excursion sensu Cramer &
Saltzman (2005, 2007), there are coupled, positive
excursions in d
13
C and d
18
O (Bickert et al. 1997;
Azmy et al. 1998) during a regression (cf. Loydell
1998). The Ireviken excursion itself occurred at a
time of climatic amelioration (Cramer & Saltzman
2005, 2007). It witnessed a positive d
13
C excursion
at globally high sea-level (Cramer & Saltzman
2005, 2007). This positive d
13
C excursion was
accompanied by a slight negative d
18
O excursion,
which may indicate warming (Bickert et al. 1997;
Azmy et al. 1998). The decoupling of isotopic and
glacioeustatic co-variation in the early Wenlock
(Fig. 2) marks a genetic change in the Earth
system. It is unlike any interval in the EPI itself,
and accompanied climatic amelioration witnessed
in the development of extensive limestone reefs
(e.g. Copper 1994; Brunton et al. 1998).
Most recent work on Early Palaeozoic
glaciations has focused on the Hirnantian (e.g.
Sutcliffe et al. 2001; Hermann et al. 2004a,b;
Armstrong et al. 2005; Le Heron et al. 2005)
since Brenchley et al. (1994) argued for a short-
lived Late Ordovician glaciation. This may partly
reflect its coincidence with a mass extinction at a
major stratigraphic division (e.g. Chen et al. 2000,
2005), as well as the deposition of major hydro-
carbon source rocks in overlying strata (Lu
¨ning
et al. 2000; Berner 2003). However, we have used
the criteria of Brenchley et al. (1994) to recognize
six more glacial maxima in the EPI. Namely, that
sequence stratigraphic evidence for large, global
lowstands, coincident with positive d
13
C and d
18
O
excursions and evidence of ice formation, indicates
a glacial maximum (Table 1; Fig. 2).
Within the EPI, individual glacial maxima
were separated by warmer intervals (e.g. Fortey &
Cocks 2005), and recent research has highlighted
clear variability in the Silurian palaeoenvironment
(e.g. Jeppsson 1990; Aldridge et al. 1993; Bickert
et al. 1997; Johnson 2006; Calner & Eriksson
2006). The frequent evidence of ice formation and
rapid, third-order variation in eustatic sea-level
(e.g. Azmy et al. 1998; Caputo 1998; Loydell
1998; Hamoumi 1999; Ghienne 2003; Nielsen
2003a,b; Johnson 2006) suggests that ice-sheets
may have dynamically expanded and retreated
during the EPI.
Glacial maxima in the EPI
The seven glacial maxima of the EPI have been
either assigned to existing named events or given
stratigraphically descriptive names based on their
nature. Thus, some are referred to as regressions
and others are called glaciations, depending on the
weight of evidence. We deal with each glacial
maximum in turn below.
The Guttenberg regression (GICE ) as defined in
Table 1 and Figure 2 is named after the Guttenberg
Limestone Member of the Decorah Fm in the Upper
Mississippi Valley, USA, where the positive
d
13
C excursion was first recognized (Hatch et al.
1987). This carbon isotope excursion has sub-
sequently been recognized elsewhere and is con-
sidered to represent a global event (e.g. Patzkowsky
et al. 1997; Ainsaar et al. 1999; Kaljo et al. 2004).
It was accompanied by a synchronous positive
d
18
O excursion of earliest clingani graptolite Zone
age (Shields et al. 2003; Tobin et al. 2005). The
high-resolution chemostratigraphy of Ludvigson
et al. (2004) shows that the GICE occurred after
three other smaller positive d
13
C excursions, and
that it took place in the P. tenuis conodont Zone
and americanus graptolite Zone (equivalent to the
earliest clingani Zone). In the Katian GSSP (Black
Knob Ridge, Oklahoma, USA), there is a seemingly
synchronous d
13
C excursion just above the Sand-
bian–Katian boundary (Goldman et al. 2005).
This is coincident with the late Keila age
regression noted by Kaljo et al. (2003) and Nielsen
(2003a); and Ludvigson et al. (2004) also noted
that the GICE occurred in a time of stratigraphical
downlap. Due to the lack of well-dated glacial
deposits in this interval (Table 2), there is no unam-
biguous link with ice formation, but it may represent
a period of cooling as continental ice-sheets were
beginning to expand (cf. Patzkowsky et al. 1997).
The early Rakvere regression (ERR) is named
after the stage in Estonia, where it is recognized
in the carbon isotope and sequence stratigraphic
records (Table 1; Fig. 2). This regression took
place within the latest clingani graptolite Zone
(Nielsen 2003a), equivalent to the A. superbus con-
odont Zone (No
˜lvak et al. 2006). Though there is
evidence for glacial erosion at this time (Table 2),
there is no evidence for extensive tillite formation.
Fig. 3. (Opposite) Correlation of marine anoxia between UK depositional basins based on the widespread occurrence
of graptolite-bearing anoxic mudrock facies, with inset schematic showing the differing depositional
environments associated with this facies. Stratigraphies and inset compiled from original work of Toghill (1968);
Cave (1979); Baker (1981); White et al. (1991). Davies et al. (1997, 2003); Pratt et al. (1995); Zalasiewicz et al.
(1995); Pore˛bska & Sawłowicz (1997); Schofield et al. (2004); and Verniers & Vandenbroucke (2006). Note that
the Hirnantian and Llandovery oxic– anoxic stratigraphy of the Welsh Basin bears close resemblance to the
Lake District and Howgill Fells succession of Northern England (Rickards 1970; Hutt 1974; Rickards &
Woodcock 2005).
BLACK SHALES IN THE EARLY PALAEOZOIC ICEHOUSE 129

Table 1. Name, age and evidence for each of the seven glacial maxima in the EPI; d
13
C
PDB
¼most positive value of carbon isotopes in positive excursion based on
regional isotopic compilations;
Dd
13
C
PDB
¼total change in carbon isotope values during positive isotope excursions based on regional isotopic compilations;
d
18
O
PDB
¼most positive value of oxygen isotopes in positive excursion recorded in brachiopod shells;
Dd
18
O
PDB
¼total change in oxygen isotope values during
positive isotope excursions recorded in brachiopod shells; oxygen isotope data taken from Tobin et al. (2005) for the Guttenberg regression, from Brenchley et al.
(1994) for the Hirnantian Glaciation in the Baltic region, and from Azmy et al. (1998) for all Silurian events; all other data compiled from Table 2, Figure 2, Kaljo
et al. (2007), and references therein
Glacial maxima Evidence for ice Dd
13
C
PDB
d
13
C
PDB
Dd
18
O
PDB
d
18
O
PDB
Extent of regression Timing
Guttenberg
regression
poorly dated glacial-erosive
features in North Africa
þ1.0‰ þ1.7‰ þ2.3‰ 22.7‰ medium early caudatus graptolite Zone
early Rakvere
regression
poorly dated glacial-
erosive features in
North Africa
þ0.8‰ þ1.8‰ – – medium late clingani graptolite Zone
early Ashgill
regression
poorly dated Ashgill tillites
and glacial erosive features
in North Africa
þ1.2‰ þ2.0‰ – – large end linearis graptolite Zone
Hirnantian
glaciation
Pan-Gondwanan tillites &
diamictites containing
Hirnantia fauna
þ4.0‰ þ4.3‰ þ4.2‰ 0‰ large extraordinarius – early
persculptus graptolite Zones
early Aeronian
glaciation
gregarius Zone diamictite
in South America
þ2.0‰ þ3.0‰ þ0.6‰ 24.5‰ small gregarius (?magnus) graptolite Zone
sedgwickii s.l.
glaciation
well-dated diamictites
in South America
þ0.5‰ þ2.0‰ þ0.7‰ 24.4‰ medium sedgwickii graptolite Zone
late Telychian
glaciation
well-dated diamictites
in South America
þ1.5‰* þ2‰* þ1.0‰* 24.6‰* large ?insectus – lapworthi graptolite Zones
*These excursions occur in the late Telychian significantly preceding the Ireviken excursion sensu Cramer & Saltzman (2005, 2007).
A. A. PAGE ET AL.130

This, along with its expression in the sea-level and
isotopic record, may indicate that it was a relatively
small, short-lived event.
The early Ashgill regression (EAR) appears to be
a more significant event based on both its sea-level
and isotopic record (Table 1; Fig. 2). Nielsen
(2003a) noted significant regression in the early
Ashgill complanatus graptolite Zone. This is
synchronous with regression and a positive d
13
C
excursion noted by Kaljo et al. (2004) at the basal
Pirgu stage in Estonia (No
˜lvak et al. 2006). There
is good evidence for glacial sediments being depos-
ited in the Ashgill of North Africa (which was close
to the palaeomagnetic South Pole). However, this
regression cannot be tied to any one particular high-
latitude event, due to imprecision in the biostrati-
graphy of these successions (Table 2).
The Hirnantian glaciation (HICE)occursastwo
pulses of glaciation within the extraordinarius–
persculptus graptolite Zones (Sutcliffe et al. 2000).
It represents the glacial maximum of the EPI and
has received extensive study. It is well constrained
and clearly globally extensive. The extent of the
eustatic and isotopic variations associated with this
are shown in Table 1 and Figure 2, with more
detailed treatment of this event being found in
Brenchley et al. (1994, 2003), Marshall et al.
(1997), Sutcliffe et al. (2001), and Armstrong (this
volume).
The early Aeronian glaciation (EAGL) has a sig-
nificant expression in marine carbon isotope values,
but this may also represent an increased effect of
carbonate weathering relative to the Ordovician gla-
ciations. There is clear evidence for ice formation
during the gregarius graptolite Zone (Tables 1 &
2), but this is a long interval, which can be subdi-
vided into the triangulatus,magnus and argenteus
zones (see Hutt 1974). Precisely how ice formation
in this event correlates with sea-level is unclear: the
Rhuddanian–Aeronian boundary regression seen in
the Ross & Ross (1996) sea-level curve is not recog-
nized in other records, which show marked
regressions at or around the magnus –argenteus
graptolite Zone boundary (Johnson et al. 1991;
Loydell 1998; Johnson 2006). Azmy et al. (1998)
show the onset of a positive d
13
C and d
18
O excur-
sion in the triangulatus Zone, but do not present
data for the magnus and argenteus zones. Kaljo
et al. (2003) illustrate a longer d
13
C excursion,
which both they and Johnson (2006) correlate to
an early Aeronian glaciation.
The sedgwickii graptolite Zone glaciation
(SEDG) is discussed at length below. Loydell
(1998) and Johnson (2006) both show major
regressions at this time. Azmy et al. (1998) and
Johnson (2006) argued for glaciations based on
isotopic data, which we note correspond to well-
dated diamictites (Table 2).
The late Telychian glaciation (LTG) is the final
glacial maximum of the EPI. As noted above, the
final pulse of diamictite deposition in South
America can be assigned to the late Telychian,
though cannot be constrained to any particular
zone (Table 2). The late Telychian is coincident
with regressions: Figure 2 illustrates a rapid sea-
level fall in the late insectus graptolite Zone (see
also Appendix), while Loydell (1998) shows a
rapid regression in the lapworthi graptolite Zone
with a lowstand throughout the lapworthi –insectus
interval, also recognized in SW Siberia by Yolkin
et al. (1997). During this interval, Azmy et al.
(1998) show a positive d
13
C and d
18
O excursion,
beginning in the crenulata graptolite Zone and
reaching a maximum in the centrifugus graptolite
Zone. Likewise, Kaljo et al. (1998, 2003) show
that the onset of this positive d
13
C transition
occurred in the late Llandovery, with subsequent
studies placing this in the Pt. amorphognathoides
conodont Zone in Estonia (Kaljo & Martma 2006)
and possibly towards its base (see Cramer &
Saltzman 2005). As such, we suggest this glaciation
occurred within the lapworthi–insectus graptolite
Zone interval.
Oxic– anoxic stratigraphy and sea-level
in the EPI
Correlating the UK oxic–anoxic stratigraphy
against glacioeustatic sea-level curves reveals a
repeated relation between black shale deposition
and deglacial transgressions throughout the EPI
(Figs 2 and 3). Conversely, glaciations themselves
correspond to well-oxygenated deep waters. Com-
parison with deep-water successions influenced by
the Rheic and Palaeotethys oceans suggests these
oxic–anoxic transitions may have had global
extent. The glacioeustatic sea-level and oxic–
anoxic stratigraphy are discussed briefly below,
before the post-Hirnantian transgression and sedg-
wickii Zone regressions are dealt with in detail.
All the oxic to anoxic transitions of the EPI rep-
resent maximum flooding surface transgressive
black shales (sensu Wignall & Maynard 1993;
Wignall 1994). They are regionally extensive and
post-date glacial maxima (Fig. 2; Woodcock et al.
1996; Armstrong et al. 2005). The GICE regression
precedes the development of extensive anoxia
within the Welsh Basin during the clingani grapto-
lite Zone. At this time, widespread, organic carbon-
rich and phosphatic black shales, such as the Nod
Glas and Cwm-yr-Eglwys Mudstone Fms (Cave &
Dixon 1993; Davies et al. 2003), were laid down
upon oxic mudrocks and limestones, such as the
Carswyn and Penyraber Mudstone Fms (Fig. 2;
Pratt et al. 1995; Davies et al. 2003). In the
BLACK SHALES IN THE EARLY PALAEOZOIC ICEHOUSE 131

Platteville–Decorah formations of eastern Iowa
(USA), the Guttenberg Member is brown shale
(Ludvigson et al. 2004). This occurs above a well-
laminated shale and widespread phosphatic bed
of the Spects Ferry Member and below the well-la-
minated shales and blackened/phosphatic
hardgrounds of the Ion Member (Ludvigson et al.
2004). Likewise, pyritic graptolite shales of cling-
ani Zone age Nakkholmen Fm in the Oslo area
(Norway) overlie limestones, as do coeval grapto-
lite shales in the lower Mossen Fm of Va
¨stergo
¨tland
(Nielsen 2003a). This represents an oxic – anoxic
transition that reflects a profound drowning event
(Nielsen 2003a,b).
The early Ashgill regression preceded the
anceps graptolite Zone transgression, with the UK
record witnessing simultaneous intervals of black
shale deposition in an otherwise oxic succession
(Fig. 3). For example, the ‘Red Vein’ in Wales, a
thin unit of anoxic mudstones bearing graptolites
of probable anceps graptolite Zone age (e.g.
Schofield et al. 2004), appears synchronous with
the anceps bands in Scotland (Williams 1982).
The Hirnantian glaciation, with an acme in the
extraordinarius graptolite Zone (Sutcliffe et al.
2000), preceded the deposition of globally exten-
sive transgressive black shales in the persculptus –
acuminatus graptolite zones (Fig. 4 & text below).
Likewise, the sedgwickii graptolite Zone glacial
event is characterized by the deposition of oxic
facies during a lowstand, which is sandwiched
between transgressive black shales (Fig. 5 & text
below). The convolutus graptolite Zone represents
a major transgression and global highstand follow-
ing the early Aeronian glaciation, with widespread
black shale deposition noted in Loydell (1998),
synchronous with the UK convolutus bands of
graptolitic shale (Fig. 3). Similarly, the deglacial
transgression in the latest Telychian is characterized
by the onset of anoxia and the deposition of marine
black shales in the centrifugus Zone in the UK.
These include the Builth Mudstones in the Welsh
Basin (Woodcock et al. 1996; Zalasiewicz &
Williams 1999) and the Brathay Mudstones in
northern England (Rickards 1970; Rickards &
Woodcock 2005). These were both deposited on
essentially oxic, late Llandovery successions
(Davies et al. 1997; Rickards & Woodcock 2005).
In Baltoscandia, this event also sees the deposition
of graptolitic shales on greenish-grey marlstones in
the Ohesaare core from Estonia (Loydell et al.
1998) and on oolitic limestone/grey-green shale
interbeds in Bornholm (Bjerreskov 1975; AAP
unpublished observations April 2006). Similarly,
Lu
¨ning et al. (2005) noted the deposition of ‘hot
shales’ on the North African/Arabian margin
during the centrifugus to firmus graptolite zones.
The relation between deglaciation, transgression
and anoxia in the late Telychian–early Wenlock
has been reviewed in depth by Cramer and Saltzman
(2007). Further examples of transgressive black
shales deposited after the Llandovery glaciations
may be found in Loydell (1998).
During the EPI, the deposition of bioturbated
facies, representing conditions of deep-marine oxy-
genation, occurred during regressions and glacial
maxima (Brenchley 1988; Loydell 1998; Fig. 2).
The Hirnantian and sedgwickii Zone glacial
maxima correspond to intervals of grey shale and
deposition of mottled (i.e. bioturbated) mudstones
(Figs 4 & 5). Likewise, the positive carbon
isotope excursion in the mid-gregarius graptolite
Zone that marks the early Aeronian glaciation
may correspond to the onset of oxic deposition in
the mid-magnus graptolite Zone of the Welsh
Basin (Fig. 3). In Black Knob Ridge, Oklahoma,
the maximum d
13
C excursion corresponding to
GICE occurs in an interval with extremely dimin-
ished C
org
content in an otherwise organic-rich
sequence (Goldman et al. 2005). This perhaps
represents an interval of increased deep-water ven-
tilation (cf. Ludvigson et al. 2004). The early
Rakvere regression (latest clingani graptolite
Zone) may possibly correlate with the transition
from black shales to limestone at Whitland, South
Wales. Also, the Fja
¨cka Shales of Sweden are
deposited on the well-oxygenated facies of the
Slandrom Limestone of Sweden and Bestorp Lime-
stone of Va
¨stergo
¨tland, representing transgressive
black shales deposited after the early Rakvere
event (Ma
¨nnil & Meidla 1994).
Transgressive anoxia: post-Hirnantian
glaciation oceanic anoxic event
The late persculptus and acuminatus graptolite
Zones are characterized by deposition of globally
extensive transgressive black shales (Fig. 4)
immediately following the extraordinarius–early
persculptus graptolite Zone acme of the Hirnantian
glaciation (Sutcliffe et al. 2000, 2001), suggesting
a fundamentally deglacial origin for the onset of
global marine anoxia. This followed the enhanced
deep ocean circulation and oxygenation that
characterized the Hirnantian glaciation (Brenchley
1988; Brenchley et al. 1994; Armstrong & Coe
1997). High-palaeolatitude sedimentary succes-
sions typically consist of Hirnantian glacial
deposits immediately overlain by black shales
(e.g. Sutcliffe et al. 2001; Armstrong et al. 2005,
2006). Low-palaeolatitude settings see unambigu-
ously oxic facies such as deep-water bioturbated
mudstones overlain by deglacial black shales (e.g.
Mu 1988; Armstrong & Coe 1997; Davies et al.
1997; Chen et al. 2000, 2005; Verniers &
A. A. PAGE ET AL.132

Vandenbroucke 2006). Shallow-water shelly
faunas in low- to mid-palaeolatitudes may also be
buried below persculptus graptolite Zone black
shales (e.g. Bjerreskov 1975; Mu 1988; Davies
et al. 1997; Chen et al. 2000, 2005), though
some shallow successions in the palaeotropics
may see limestone deposition going on uninter-
rupted (e.g. Barnes & Bolton 1988). These degla-
cial black shales are widely palaeogeographically
distributed and represent a global event (Fig. 4)
perhaps comparable to the Mesozoic oceanic
anoxic events (cf. Cohen et al. 2004).
Deposition of regionally extensive black shales
on maximum flooding surfaces (sensu Wignall &
Maynard 1993; Wignall 1994) in the persculptus
graptolite Zone provides strong evidence for their
onset in the end Ordovician–early Silurian trans-
gression (Ross & Ross 1995, 1996; Loydell 1998;
Lu
¨ning et al. 2000; Nielsen 2003a). However,
precise biostratigraphic dating of high-latitude
black shales is hindered by the relative scarcity of
graptolites in these settings (Skevington 1974;
Zalasiewicz 2001) and the prevalence of non-
diagnostic taxa with long ranges (Lu
¨ning et al.
2000). Nevertheless, where high-latitude, post-
glacial black shales contain a sufficient fauna to
permit dating, the onset of anoxia can be assigned
to the persculptus graptolite Zone: e.g. the black
shales of the Cedarberg Fm, South Africa, and the
Don Braulio Fm, Argentina (Sutcliffe et al. 2000,
2001); Batra Fm, Jordan (Armstrong et al. 2005);
and Murzuq Basin, Libya (Lu
¨ning et al. 2000).
Though Lu
¨ning et al. (2006) noted that the base
of the Batra Fm is diachronous, with its lower
member yielding an acuminatus graptolite Zone
fauna towards the (present-day) North (Armstrong
et al. 2005, 2006), this is neither inconsistent with
the onset of anoxia occurring in the persculptus
graptolite Zone, nor is it inconsistent with depo-
sition of the Batra Fm as a maximum flooding
surface black shale. As the transgression continued
into the early Silurian, the oxygen minimum zone
would have shoaled further up the shelf (Armstrong
et al. 2006). The formation of early Silurian black
shales, such as at the base of the Qusaiba Shale,
Saudi Arabia (Aoudeh & Al-Hajri 1995), is
Fig. 4. Schematic lithographic logs showing the onset of transgressive anoxia in different settings during the
persculptus graptolite Zone anoxic event. Palaeogeographical reconstruction showing the position of continents and
associated terranes in the Hirnantian, annotated with localities where black shale is deposited on a maximum flooding
surface: 1. Bavnega
˚rd Well, Bornholm, Denmark (Bjerreskov 1975); 2. Dob’s Linn, Scotland UK (Toghill 1968;
Armstrong & Coe 1997; Verniers & Vandenbroucke 2006); 3. Fenxiang, Yingang, Yangtze region, southern China (Mu
1988; Chen et al. 2000, 2005); 4. Cwmere Fm, central Welsh Basin, UK (Woodcock et al. 1996; Davies et al. 1997);
5. Murzuq Basin, Libya (Lu
¨ning et al. 2000); 6. Batra Fm, Jordan (Armstrong et al. 2005). Palaeogeographic
reconstructions mainly after Torsvik et al. (1998). The relative position of Gondwana, Armorica, Baltica, Avalonia,
Laurentia and Siberia are largely unmodified.
BLACK SHALES IN THE EARLY PALAEOZOIC ICEHOUSE 133

evidence that oceanic anoxia and deposition of
maximum flooding surface black shales continued
as the transgression continued.
The influx of deglacial meltwater in the oceans
of the persculptus graptolite Zone may have been
critical to the onset of transgressive anoxia: it may
have increased marine stratification through the for-
mation of low-salinity surface waters and, by pro-
viding a source of nutrients via continental
weathering, stimulated marine productivity. This
may be analogous to the formation of sapropels in
the Neogene Mediterranean Basin (Rohling &
Gieskes 1989; Rohling 1994; Scrivner et al.
2004). Buoyant, low-salinity surface waters,
strengthening the pycnocline in the deglacial
Hirnantian Ocean, may have precluded deep-water
thermohaline circulation to sufficiently maintain a
well-oxygenated sea-floor. Periglacial outwash
may have carried sufficient nutrients to fuel the
deposition of the ‘hot shales’ of North Africa and
Arabia (cf. Meybeck 1982), which are characterized
by a total organic-carbon content of up to 17%
(Lu
¨ning et al. 2000), well above that found in
normal black shales.
Some authors have declared that ‘hot shale’
deposition may be a result of upwelling (Lu
¨ning
et al. 2000, 2005, 2006), but this disaccords with
both their widespread, synchronous deposition and
with GCMs of Hirnantian circulation (Hermann
et al. 2003, 2004a,b). Upwelling is a regionally
localized phenomenon and the oxygen minimum
zone associated with upwelling zones is only
stable on decadal timescales (Wignall 1994). More-
over, meridional and monsoonal coastal upwelling
is restricted to low to mid-latitudes (Parrish 1982),
so are unlikely to apply to the ‘hot shales’, which
occur at high palaeolatitudes. Meanwhile,
end-Ordovician continental configuration is
inconsistent with widespread zonal coastal upwel-
ling (Armstrong et al. 2006). Zonal upwelling
occurs when north or south continental margins
lie adjacent to the major zonal wind systems
(Parrish 1982). Comparing modern palaeogeogra-
phical reconstructions (Scotese & McKerrow
1991; Cocks & Torsvik 2002) with the high-latitude
zonal wind predicted in the late Ordovician atmos-
pheric simulations of Parrish (1982) shows that
major winds were primarily orthogonal to the
Fig. 5. Schematic lithographic logs showing evidence of regressive oxygenation during the sedgwickii graptolite
Zone glaciation. Palaeogeographical reconstruction showing the position of continents and associated terranes in
the mid-Llandovery, showing global event of deep-water oxygenation: 1. Girvan Group, Scotland, UK (Floyd &
Williams 2003); 2. Ølea
˚, Bornholm, Denmark (Bjerreskov 1975); 3. Western Iberian Cordillera, NE Spain
(Gutie
´rrez-Marco & S
ˇtorch 1998); 4. Qusaiba Shale, Qalibah Fm, Saudi Arabia (Miller & Melvin 2005; AAP/
JAZ/MW unpublished observations). Palaeogeographic reconstructions again mainly after Torsvik et al. (1998),
with amendments as noted in Figure 4. Also, by the beginning of the Silurian, the amalgamation of some
Kazakh crustal terranes probably led to the formation of a substantial landmass north of Tarim and South China
(Koren et al. 2003). The position of Annamia close to South China and the Karakum– Tajik plate is mainly based on
the strong affinities of the late Silurian (Ludlow– Pridoli) brachiopod faunas (Rong et al. 1995; Thong-Dzuy et al.
2001).
A. A. PAGE ET AL.134

Gondwanan margin, making widespread zonal
upwelling untenable. This is borne out by recent,
more sophisticated GCMs for the late Ordovician,
that show predominantly onshore ocean currents
around the North African and Arabian margins of
Gondwana (Hermann et al. 2003, 2004a,b). As
such, upwelling alone can neither account for the
simultaneous onset of globally widespread anoxia
in the persculptus graptolite Zone nor for the
widespread high-latitude occurrence of the most
organic-rich shales.
Furthermore, deglacial melting would most
likely be to the detriment of the increased thermo-
haline circulation needed to sustain upwelling. In
the Quaternary of the North Atlantic, intervals of
meltwater outwash are associated with increased
ocean stratification and more sluggish deep-water
circulation (Rahmstorf 2002). Instead, an alterna-
tive nutrient source may lie in increased continen-
tal weathering during deglaciation, providing a
major source of both major and trace nutrients
(Broecker 1982; Meybeck 1982) and stimulating
productivity (as modelled by Kump & Arthur
1999). Given that the locus of glacial melting is
focused on the Gondwana palaeocontinent, this
nutrient source could explain the relatively
increased organic content of high-palaeolatitude
black shales compared to those in low palaeolati-
tude black shales. This would leave anoxia at low
palaeolatitudes as a consequence of increased
ocean stratification and decreased thermohaline cir-
culation due to high-palaeolatitude ice melting.
This, coupled with sediment starvation (cf.
Wignall 1991), would lead to an expanded oxygen-
minimum zone. Hence, periglacial run-off provides
a mechanism for anoxia that would have simul-
taneously increased global seawater stratification
as well as stimulating productivity and export pro-
duction (see schematic in Fig. 6). This highlights
the fundamentally deglacial nature of transgressive
black shales in the EPI.
Regressive oxygenation: the sedgwickii
graptolite Zone event
The sedgwickii graptolite Zone is marked by global
regression of plausibly glacial origin during which
Fig. 6. (a) Summary cartoon showing end-member transgressive and regressive oceansin the EPI, outlining a model in
which transgressive black shale deposition may serve as a negative-feedback mechanism modulating glacioeustasy
(b) Graphs showing the postulated relationship between CO
2
temperature and sea-level, given that intervals of
transgressive black shale deposition may draw-down significant atmospheric CO
2
.
BLACK SHALES IN THE EARLY PALAEOZOIC ICEHOUSE 135

sediments were deposited in well-oxygenated deep
water overlying earlier black shales (Table 1;
Figs 2 & 5). Graptolites are rarely preserved in exten-
sively bioturbated facies, though graptolites could
clearly exist in well-oxygenated waters (Armstrong
& Coe 1997; Floyd & Williams 2003; AAP/JAZ/
MW unpublished observations). Therefore, biotur-
bated graptolite-free intervals in otherwise continu-
ally graptolitic successions may be taken as
evidence of increased sedimentary ventilation, seen
in both high- and low-latitude successions in the
Iapetus, Rheic and Palaeotethys Oceans (Fig. 5 &
references therein). It is clear that the boundary
between the preceding convolutus graptolite Zone
and the sedgwickii graptolite Zone is characterized
by the development of oxic strata, and that wide-
spread anoxia is again developed by the onset of
the guerichi graptolite Zone (Loydell 1998).
However, the sedgwickii graptolite Zone also
contains a distinctive interval of black shale,
which is only seen in certain successions from the
Iapetus and Rheic Oceans. In UK successions in
Wales and Dob’s Linn, this thin black shale is sand-
wiched between a sequence of grey shales and
mottled mudstones (e.g. Toghill 1968; Cave
1979). The sedgwickii Zone of the Kullatorp core
(Va
¨stergo
¨tland, Sweden) comprises greyish-green
mudstones in a succession of otherwise graptolitic
black shales. However, the Silvberg and
Gullera
˚sen-Sanden sections of Dalarna contain a
finely laminated shale yielding St. sedgwickii
(Loydell 1998). In Bornholm, Denmark, however,
there is an abrupt transition between the highly
graptolitic black shale of the cometa Band that
characterizes the top of the convolutus graptolite
Zone and a heavily bioturbated, non-graptolitic
silty, grey mudstone (Bjerreskov 1975; AAP
unpublished observations on Ølea
˚section and col-
lections in the Geological Museum, University of
Copenhagen). The sedgwickii shale band in the
Welsh Basin is generally thought to represent a
transgressive black shale (Cave 1979; Baker 1981;
Davies & Waters 1995; Woodcock et al. 1996).
However, it is unlike the transgressive black
shales of the persculptus graptolite Zone and early
Wenlock (Woodcock et al. 1996; Lu
¨ning et al.
2005; Armstrong et al. 2005, 2006) in which
organic-carbon burial is more profoundly
expressed at high palaeolatitudes. Instead, this
brief black shale event in the sedgwickii Zone is
only observed at certain low-latitude settings.
There is no major graptolitic shale burial associated
with the sedgwickii Zone age strata in the
Qusaiba Shale, Saudi Arabia (Zalasiewicz et al.
2007). This may reflect increased isolation
of semi-enclosed basins at low sea-level,
suggesting local rather than global significance
(e.g. Loydell 1994).
There is good evidence of a regression within
the lithofacies deposited during the sedgwickii grap-
tolite Zone, and no graptolites have been recovered
from the low-latitude successions of the Western
Iberian Cordillera, Spain (Gutie
´rrez-Marco &
S
ˇtorch 1998), the Prague Basin, Czech Republic
(S
ˇtorch 1986, 1994), and the Canadian Arctic
(Melchin 1989). In Girvan, southern Scotland, UK,
the late sedgwickii Zone contains shallow-water
brachiopods and deposits with hummocky cross-
stratification in the Lower Camregan Grits Fm
overlying the black shales of the Pencleuch Shale
Fm (Floyd & Williams 2003), and in Spengill,
Howgill Fells, UK, the sedgwickii graptolite Zone
contains the only occurrence of limestones and
grits in its Llandovery succession (Rickards 1970).
The evidence for a regression in UK deep-water
strata may correlate with the regressions recognized
in the sedgwickii s.l. graptolite Zone in sequence
stratigraphic analyses of shallow-water facies in
Baltoscandia, North America and Siberia (Johnson
1996; Ross & Ross 1996; Yolkin et al. 1997), and
coincides with the formation of diamictite in
Gondwana (Caputo 1998); whilst the overlying
guerichi graptolite Zone itself appears to mark the
onset of extensive marine anoxia (Loydell 1998),
which may be linked to a transgression (Fig. 2).
Increased oxygenation of the marine realm in the
sedgwickii graptolite Zone may be a consequence of
high-latitude ice formation stimulating more rapid
thermohaline circulation, as is seen in the modern-
day Atlantic (Rahmstorf 2002). The sedgwickii
Zone regression is coincident with diamictite depo-
sition in South America (Caputo 1998; Table 1;
Fig. 2), suggesting glacioeustatic control. The for-
mation of marine ice would have resulted in brine
rejection, creating cold, dense waters at high lati-
tudes. On sinking, these may have driven a more
vigorous deep-water circulation, providing a
greater flux of oxygen to the deep oceans (see
Fig. 6), consistent with the well-ventilated deep-
water facies seen in glacial maxima during the EPI
(Fig. 2; Brenchley 1988). The global cessation of
anoxic facies at the end of the convolutus graptolite
Zone and their return in the guerichi graptolite Zone
reflects a third-order change in depositional style,
consistent with a glacioeustatic control on oceanic
redox state (Church & Coe 2003).
A simple model for carbon burial and
glacioeustasy in the EPI
During the EPI, there was a fundamental link
between glacioeustatic sea-level change and the
burial of organic carbon in deglacially transgressive
black shales, which may have represented a nega-
tive feedback mechanism that served to stabilize
A. A. PAGE ET AL.136

climate (Fig. 6), and prevented the onset of runaway
greenhouse conditions. Given that sea-level may
represent a proxy for atmospheric temperature,
which itself is a function of pCO
2
, the deposition
of globally extensive black shales on maximum
flooding surfaces may have served to slow the
initial warming after the glacial maxima by
drawing down CO
2
from the ocean–atmosphere
system. This would sustain the EPI and prevent
onset of greenhouse conditions that characterized
most of the Early Palaeozoic (Frakes et al. 1992;
Gibbs et al. 2000; Montan
˜ez 2002; Church & Coe
2003).
This model for deposition of black shales due to
periglacial meltwater increasing both productivity
and ocean stratification predicates that oceanic
anoxia was intimately linked to glacioeustasy.
Black shale deposition in transgressions may have
been significant to influence cooling, and therefore
a regression, by drawing down atmospheric
carbon (see the schematic graph in Fig. 6). Like-
wise, the onset of well-oxygenated oceans due to
brine rejection driving deep-water circulation may
have served to prevent organic carbon burial to
sustain fully glacial conditions, which might have
led to another transgression. With global marine
oxygenation linked to ice formation and melting,
we infer a strong link between organic carbon
burial in black shales, sea level and atmospheric
CO
2
during the EPI. This model now needs to be
rigorously tested against both the sedimentary
record of the EPI and by developing theoretical
models of the carbon budget.
Black shale deposition and the EPI
carbon cycle
Though we have clearly shown a strong link
between black shale deposition and deglacial trans-
gressions, relating anoxia and carbon-burial flux is
not straightforward. Models of the carbon cycle
show a significant increase in carbon burial as
black shales close to the Ordovician–Silurian
boundary (Fig. 1c). Meanwhile, Ronov et al.
(1980) estimated that the amount of organic
carbon buried in Ordovician and Silurian black
shales is comparable to that in Permo-
Carboniferous strata, a time when the organic reser-
voir may have exerted a significant influence on
atmospheric CO
2
(e.g. Bruckschen et al. 1999).
As the EPI corresponds to a major low in shelf-
carbonate deposition, organic carbon burial in
black shales may have been critical to regulating
the carbon cycle (Fig. 1; Patzkowsky et al. 1997).
The increase in atmospheric O
2
during this interval
is consistent with increased organic-carbon burial
(Fig. 1a, c; cf. Berner 2003; Catling & Clare
2005), empirically linking black shale deposition
and atmospheric change. Though we have shown
that black shale burial occurs in transgressions,
the lateral extent of black shale is unclear, making
it hard to estimate the carbon burial flux from
rock volume.
The oxygenation–depth profile of EPI oceans is
largely unknown. Oxic–anoxic transitions occur in
open-ocean settings, demonstrating that this is not a
phenomenon exclusive to restricted basins or epi-
continental seas (cf. Pore˛bska & Sawłowicz
1997). Similarly, anoxia may be developed in
shelf settings relatively close to the storm wave-
base, such as in the Qusaiba Shale of Saudi
Arabia, where anoxic or dysoxic facies alternate
with bioturbated mudstones and facies yielding
benthic faunas (Miller & Melvin 2005). Precisely
how shallow anoxic conditions may extend in the
EPI is unclear (cf. Pancost et al. 1998). Whether
these transgressive anoxic events represent shoaling
of an expanded oxygen minimum zone onto the
shelf, or whether there was widespread whole
oceanic anoxia in the EPI, remains uncertain.
We concur with Cramer & Saltzman (2005,
2007) that the deposition of organic carbon in epi-
continental black shales cannot account for positive
d
13
C excursions in the Silurian. As they noted,
widespread deposition of shales in epicontinental
seas does not coincide with these excursions.
Neither do our data show evidence for increased
carbon burial in deep-water settings with strong
oceanic influence during any of the positive d
13
C
excursions (Table 1; Fig. 2). We also contest the
suggestion of upwelling and increased carbon
burial as an explanation for positive d
13
C excur-
sions in the EPI (e.g. Young et al. 2005). All posi-
tive d
13
C excursions in the EPI are best explained
by increased weathering of shallow marine carbon-
ates in regressions (e.g. Kump et al. 1999; Melchin
& Holmden 2006) as noted in the Appendix. Young
et al. (2005) argued that the Guttenberg regression
was synchronous with biomarker evidence for
photic-zone anoxia based on the data of Pancost
et al. (1998). This however represents a miscorrela-
tion. The high-resolution stratigraphy of Ludvigson
et al. (2004) demonstrates that the interval of photic
zone anoxia identified by Pancost et al. (1998) cor-
responds to the Spects Ferry Member of the Platte-
ville Fm. The Spects Ferry Member underlies the
Guttenberg Limestone, preceding the EPI, whilst
the Guttenberg regression corresponds to lithologi-
cal evidence of a more-oxic interval (cf. Ludvigson
et al. 2004). As noted above, there is no direct evi-
dence for increased organic carbon preservation or
productivity during the positive d
13
C excursions
of the EPI (cf. Chen et al. 2005). Because of this,
and the inability of upwelling to explain the syn-
chronous onset of global anoxia, we attribute
BLACK SHALES IN THE EARLY PALAEOZOIC ICEHOUSE 137

anoxia to increased deglacial outwash causing
oceanic stratification (Fig. 6).
It is still unclear how black shale burial varies
with regard to the burial of other facies associated
with atmospheric CO
2
drawdown. On timescales
greater than 1000 years, carbonate burial provides
a sink for CO
2
(Elderfield 2002). Continental silicate
weathering draws down CO
2
over geological time-
scales and may have been the key driver of
changes in atmospheric CO
2
over geologically long
timescales (Holland 1978; Berner 1991; Raymo
1991; Kump et al. 1999; Cohen et al. 2004). As con-
tinental weathering is thought to increase with
greater temperatures (Berner et al. 1983; Velbel
1993; Hovius 1998), and the rate of erosion
appears to be greater in unstable (transitional) cli-
mates than in stable greenhouse or icehouse climates
(e.g. Shuster et al. 2005), one would expect a greater
CO
2
drawdown due to weathering in transgressions.
Our model for drawdown of CO
2
in transgressive
anoxia due to increased freshwater runoff from the
continents would augment this, complementing
drawdown of CO
2
to the oceans by silicate weather-
ing and limestone burial (e.g. Kump et al. 1999).
Improved constraints on the relation between weath-
ering, carbonate deposition, black shale burial and
changes in global temperature would further
improve our understanding of EPI climate.
Comparison of black shale and carbonate
burial in the EPI
Black shale burial can be compared to carbonate
deposition and changes in global temperature by
relating the oxic/anoxic stratigraphy of Figure 2
to Primo/Secundo (P/S) and Humid/Arid (H/A)
cycles. These cycles relate changes in carbonate
deposition to changes in temperature and deep-
water circulation. P/S and H/A cycles were recog-
nized in the Silurian by Jeppsson (1990, 1997) and
Bickert et al. (1997) respectively. These cycles
have recently been reviewed and synthesized by
Cramer & Saltzman (2005; 2007), who viewed
Silurian climate as alternating between two end-
member oceanographic regimes. The first regime
is characterized by cooler (P), wetter (H) climates
with high sea-levels; argillaceous limestone depo-
sition took place in shallower successions and
black shales were deposited on the continental
shelf, with oxic deep waters. The other regime is
characterized by warmer (S), more arid (A) climates
with lower sea-level; reefs formed in shallow suc-
cessions and limestones were deposited on the
shelf with black shales deposited in anoxic open
oceans. It should, however, be noted that Jeppsson
(1990, 1997) and Bickert et al. (1997) differ in
their views on where and when anoxia occur.
Bickert et al. (1997) suggested the open oceans
were anoxic throughout both H and A episodes
with anoxia shoaling onto the shelf in H episodes.
Jeppsson (1990, 1997) proposed oxic deep waters
occurred in P episodes with anoxia occurring on
the shelf due to high productivity in these episodes;
conversely, he proposed that S episodes were
characterized by well-oxygenated shelf conditions
and anoxia in the open oceans.
There is no simple relationship between the
oxic–anoxic stratigraphy of the British Isles and
either P/SorH/A cycles. In the Ordovician, Kaljo
et al. (1999; 2004) recognized humid and arid epi-
sodes in both the mid–late Caradoc, which was pre-
dominantly anoxic, and the Ashgill, which was
predominantly oxic (Fig. 3). Likewise, observations
of glacial maxima and anoxia do not accord well
with the P/S model. For example, the Spirodden
Secundo episode lasts from the persculptus grapto-
lite Zone until the argenteus graptolite Zone
(Aldridge et al. 1993), a time recognized as being
a warmer deglacial interval. This episode has a
similar duration to the anoxic conditions in the UK
(persculptus–magnus graptolite zones interval) as
shown in Figure 2. In contrast, the Malmøykalven
Secundo episode (Aldridge et al. 1993) corresponds
to the sedgwickii graptolite Zone, a period of glob-
ally extensive oxic conditions and glaciation, as dis-
cussed above. Moreover, the P/S episodes
recognized in the Llandovery (Aldridge et al.
1993) do not correspond well with glacioeustatic
sea-level changes or episodes of organic carbon
burial in black shales (Fig. 2; Loydell 1998).
There was a relatively low rate of limestone burial
during the EPI (Fig. 1b), and glacial maxima show
a clearer link between changes in oceanic redox
state than they do with either P/SorH/A cycles.
So, we suggest that ocean anoxia seems to reflect
changes in atmospheric CO
2
and temperature more
clearly than does carbonate deposition.
Discussion
The model works well for the clearly deglacial
anoxic events in the EPI, especially those in the
clingani,persculptus and centrifugus graptolite
zones (Fig. 2). Likewise, there is clear evidence
for increased deep-water ventilation and the depo-
sition of well-oxygenated mudstones in glaciations
themselves (Figs 2 & 5). However, it only
addresses how black shale deposition may serve
as a negative feedback to stabilize EPI climate
rather than providing a mechanism for the onset
and termination of ice formation in the EPI. Recog-
nizing repeated glacial maxima within the EPI,
rather than just focusing on the Hirnantian event,
significantly furthers our understanding of Early
A. A. PAGE ET AL.138

Palaeozoic climate. Characterizing the lithostrati-
graphic patterns associated with each glacial
maximum has established a consistent set of
events associated with ice-sheet formation and
retreat. The factors associated with ice formation
within individual EPI glacial maxima can be poten-
tially inferred from these ‘event stratigraphies’.
The differences in scale between each of the
glacial maxima may be used to infer which
factors were most important for ice formation in
the EPI. That is, the most important factors control-
ling ice formation are likely to be most strongly
expressed in large events such as the Hirnantian
glacial maximum, but only weakly expressed in
smaller events. Once the factors which control ice
formation in glacial maxima have been established,
it may be possible to infer what was responsible for
the onset and termination of the EPI.
We have employed stratigraphic correlation of
sub-graptolite zone resolution to recognize individ-
ual glacial maxima within the EPI and their associ-
ated lithostratigraphic changes. By combining
isotopic data with glacioeustatic curves and lithos-
tratigraphic patterns, there is considerable potential
for developing a highly resolved Ordovician –
Silurian global event stratigraphy. By comparison,
biostratigraphy alone offers comparatively poor
resolution and global coverage. The first appear-
ances of many graptolites are diachronous (cf.
Williams et al. 2003, 2004; Cooper & Sadler
2004). Likewise, the relative scarcity of graptolites
in cool waters and in shallow environments
(Skevington 1974; Finney & Berry 1997; Loydell
1998; Zalasiewicz 2001) may preclude accurate
correlation between high and low latitudes (e.g.
Lu
¨ning et al. 2000), and also between graptolite
and conodont or shelly-fossil biostratigraphies
(e.g. Mullins & Aldridge 2004). Contrastingly,
carbon isotope stratigraphies appear to show good
correlations between differing facies and palaeo-
geographical settings, allowing global correlation
(e.g. Underwood et al. 1997; Melchin & Holmden
2006; Kaljo et al. 2007). Moreover, the strong coup-
ling of both positive carbon isotope excursions with
changes in deep-water redox conditions and
glacioeustasy (Figs 2 & 6) represents a method of
correlating third-order sequence stratigraphic
changes, potentially providing a new tool for
Early Palaeozoic stratigraphy. Likewise, this
model for formation of transgressive black shales
and their role in carbon burial (Fig. 6) may extend
to other intervals. For example, the Toarcian
oceanic anoxic event in the Jurassic saw continental
run-off corresponding to the onset of anoxia, which
induced CO
2
drawdown and subsequent cooling
(Cohen et al. 2004).
In the EPI, neither the occurrence of glacial
maxima nor oxic–anoxic transitions are regularly
spaced (Fig. 2) and these events may therefore
have been externally forced. The lack of observed
cyclicity in the occurrence of glacial maxima
suggests that the negative feedback due to CO
2
drawdown in transgressive black shale deposition
was insufficient to create regular, self-sustaining
glacial–interglacial cycles. The long intervals of
anoxic marine conditions (e.g. late Caradoc, early
Silurian and late Telychian–early Wenlock) and
predominantly oxic marine conditions (e.g. early
Caradoc, Ashgill, Telychian) may suggest that
there was a second-order, possibly tectonic control
on ocean redox conditions (see also Leggett 1980;
Leggett et al. 1981). Likewise, the distribution of
the currently recognized glacial maxima may be
of second-order rather than third-order periodicity.
Such second-order forcing, possibly by the
opening and closing of oceanic gateways (e.g.
Smith & Pickering 2003; Armstrong this volume),
or drawdown of CO
2
due to increased silicate
weathering in orogenies (Kump et al. 1999), may
ultimately be responsible for the conditions that
facilitated ice formation in the EPI.
Conclusions
Recognizing the EPI as a long-lived interval revises
our understanding of Early Palaeozoic climate,
showing a long-lived icehouse in an interval pre-
viously thought to be dominated by greenhouse
conditions (cf. Brenchley et al. 1994; Gibbs et al.
2000; Montan
˜ez 2002; Church & Coe 2003;
Royer 2006). Its c. 30-million-year duration makes
it comparable to other long-lived icehouses such as
those in the Cenozoic, Permo-Carboniferous and
even the Neoproterozoic. All of these events are
characterized by waxing and waning of ice-sheets
on different timescales with potentially different
forcing and feedback mechanisms dependent on
timescale. Moreover, each of these icehouses rep-
resents a markedly different solution to the
Earth’s carbon budget, and the locus and import-
ance of organic burial varied considerably
between them (cf. Fig. 1). Though atmospheric
CO
2
levels appear to have controlled global temp-
erature over geological time (Royer 2006), each
of these icehouses is characterized by a markedly
different carbon cycle, which needs to be under-
stood in its own terms.
Though there are many uncertainties shrouding
our understanding of glaciation in this interval,
the EPI is notably different from those of the
modern oceans: it may have represented a time
when the deposition of black shales in conditions
of marine anoxia played a significant role in mediat-
ing the carbon cycle and climate.
BLACK SHALES IN THE EARLY PALAEOZOIC ICEHOUSE 139

We wish to thank David Harper and an anonymous
reviewer for their helpful and insightful comments. AAP
wishes to thank Eddie Blackett, Andrea Snelling (both
Leicester & BGS), Nigel Woodcock & Barrie Rickards
(Cambridge) and Jakob Vinther (Copenhagen) for
showing him around persculptus and sedgwickii graptolite
Zone sections in Wales, Sedbergh and Bornholm, and for
providing access to their collections from these areas.
Likewise, we thank Merrell Miller (Saudi Aramco) for
providing access to boreholes of late Aeronian–early
Telychian strata from Saudi Arabia. Cambridge Earth
Science contribution 8775.
Appendix: Choice/interpretation of
datasets and correlations
Details of individual datasets used in this study and our
interpretations of them are discussed in turn below.
Stratigraphic framework: our correlations follow those
of Cooper & Sadler (2004) for the Ordovician, and
Melchin et al. (2004) for the Silurian, although we
correlate the Llandovery–Wenlock boundary to Ireviken
datum 2 (see next paragraph). In addition, we refer to
No
˜lvak et al. (2006) for correlation of the Ordovician
timescale in Estonia to the global standard. We refer to
other stratigraphic compilations in the text as necessary
and assume that all major isotopic excursions are globally
synchronous.
The International Commission on Stratigraphy website
notes that the correlation of the Llandovery–Wenlock
boundary GSSP is ‘imprecise’ (www.stratigraphy.org).
To resolve this, the stratigraphy of the late Telychian–
early Wenlock interval has received much attention of
late (e.g. Jeppsson 1997; Loydell et al. 2003; Munnecke
et al. 2003; Mullins & Aldridge 2004; Cramer & Saltzman
2005, 2007). The absence of taxonomically identifiable
graptolites in the GSSP (Mullins & Aldridge 2004)
makes correlation to graptolite zones uncertain
(Melchin et al. 2004). However, the GSSP has a good con-
odont and microfossil stratigraphy, although the position
of the golden spike does not correspond to the base of
any particular biozone (Mabillard & Aldridge 1985;
Mullins & Aldridge 2004). Instead, recent works have cor-
related this boundary at a slightly younger level, namely
Ireviken datum 2 (e.g. Loydell et al. 2003; Calner et al.
2004). This represents the boundary between the lower
and upper Ps. bicornis conodont zonal levels (Jeppsson
1997), which is close to the base of the murchisoni grapto-
lite Zone (Loydell et al. 2003). This level has been
suggested as a correlatable level for the boundary on the
International Commission on Stratigraphy website
(www.stratigraphy.org).
Most original works on Ordovician deposits (Table 2)
correlate these strata to the British stages (cf. Fortey et al.
1995, 2000). As such, we refer to these stages throughout
this paper (the relations between the British stages and the
international stages of Cooper & Sadler [2004] are shown
in Fig. 2). However, we use the term Hirnantian sensu
Cooper & Sadler (2004) as this stage is well-defined
with good global correlation. We note that the definition
and correlation of the British Ordovician stages is not
unproblematic (Fortey et al. 1995, 2000; Cooper &
Sadler 2004). The recent placement of the GSSPs for
these stages reflects improved Ordovician biostratigraphy.
The recently named Sandbian and Katian Stages of the
Ordovician are defined on the well-correlated first appear-
ances of the graptolites Nemagraptus gracilis and Ensi-
graptus caudatus, even though the first appearance of
these graptolites is locally diachronous (cf. Williams
et al. 2003, 2004). Therefore, we feel that the historical
correlation of glacial deposits to the British stages, and
our use of the UK oxic –anoxic stratigraphy, justify our
reference to these ‘old-fashioned’ terms and we wish to
highlight that use of old terms does not necessarily
reflect the employment of outmoded correlations.
Criteria for recognizing glacial maxima: The glacial
maxima identified in the EPI (Table 1) are recognized
using an argument similar to that employed by Brenchley
et al. (1994). Namely, glacial maxima occur when rapid
regressions are accompanied by synchronous oxygen and
carbon evidence of cooling, if there are contemporaneous
glacial deposits. Glaciations may be recognized from
either the deposition of diamictites containing glaciogenic
clasts, ice-rafted debris in distal marine settings, or from
glacial erosive features (Eyles 1993). However, evidence
of ice may not necessarily be evidence of glacial
maxima, as ice-sheets may persist through interglacials.
Also, an extensive unconformity of Caradoc– Hirnantian
age persists through much of Africa and Arabia
(Destombes et al. 1985; Sutcliffe 2001), potentially
removing evidence of glaciations in this interval. Like-
wise, correlation between high-latitude glacial deposits
and equivalent low-latitude strata may be hindered by
(a) the low-abundance of graptolites in high-latitude
environments (Lu
¨ning et al. 2000; Zalasiewicz 2001),
and (b) the general absence of limestone-hosted shelly
faunas in these settings (e.g. Walker et al. 2002).
However, if the co-occurrence of ice formation with
regressions and positive d
13
C and d
18
O excursions is not
contradicted by biostratigraphic data, then it seems more
parsimonious to consider them to be related to a glaciation
rather than being caused by separate events.
Sea-level curves: The sea-level curves illustrated in
Figure 2 are based on sequence stratigraphic analyses of
shallow-water facies. These may be ‘calibrated’ to the
depths of such facies in modern environments (Ross &
Ross 1996), though such estimates may also vary accord-
ing to sediment flux or local topography (Orr 2001).
Nevertheless, other methods for estimating sea-level
change possess inherent uncertainties. Faunally derived
sea-level curves may represent changes in palaeoenviron-
ment rather than deepening per se (Orr 2001). For
example, ‘quantitative’ sea-level curves based on
conodont assemblages do not produce consistent results
in different environments (e.g. Zhang et al. 2006).
Glacial maxima within the EPI may be associated with
A. A. PAGE ET AL.140

Table 2. Evidence for ice formation during the EPI; palaeolatitudes inferred from palaeogeographical reconstructions for the Caradoc, Hirnantian and Llandovery–
Wenlock boundary, quoted to within +58N/S; stratigraphic nomenclature reflects the original author’s usage
Stage Age Location Palaeolat. Evidence of ice Evidence for age
CARADOC sensu
Fortey et al.
(1995, 2000)
‘Llandeilo-Caradoc
boundary’ Hamoumi
(1999)
Lower Ktaoua Formation,
Zagora, Morocco
808S glacial pavement with polished
surface displaying roche
moutonne
´e-like forms,
undulating surfaces, graze,
score & nail-shaped groove
joints (Beuf et al. 1971;
Hamoumi 1999)
surface forms base of L. Ktaoua
Fm, which has ?gracilis–clingani
graptolite Zone age (Destombes
et al. 1985) based on comparison
of trilobite & brachiopod fauna
with UK and Bohemia; overlies
1st Bani Formation of ‘Llandeilo’
age (Destombes et al. 1985) with
erosive contact
‘Lower Caradoc’
Hamoumi (1999)
Lower Ktaoua Formation,
Zagora, central Anti-Atlas,
Morocco
808S surface remnants at corrie heads
displaying battered & scoured
surfaces, nivation hollows and
thermokarst, on top of
glaciotectonised and jointed
glaciomarine sediments (Beuf
et al. 1971; Hamoumi 1999)
within L. Ktaoua Fm, which has
?gracilis–clingani graptolite
Zone age (Destombes et al.
1985), based on comparison of
trilobite & brachiopod fauna with
UK and Bohemia
Caradoc (Pickerill et al.
1979)
Gander Bay Tillites,
Davidsville Group,
Newfoundland, Canada
658S
†
diamictites and locally abundant
dropstones (Pickerill et al.
1979)
oldest strata in the Davidsville
Group contain conodonts of
Llanvirn–Llandeilo age;
diamictites immediately underlie
graptolitic slate of Caradoc age
(Pickerill et al. 1979)
Caradoc (Schenck &
Lane 1981)
White Rock Fm, Nova
Scotia, Canada
658S
†
marine-rafted tillite with clasts in
small lenses and dropstone
fabrics in varvites (Schenck &
Lane 1981), quartz
meta-arenites showing
polished striations and groves
on facets (Schenck 1972)
tillite overlain by ‘poorly preserved,
limited [graptolite] fauna of
Caradoc or younger age’
(Schenck 1972; Schenck & Lane
1981)
CARADOC or
ASHGILL
‘middle Caradoc– Ashgill
boundary’ Hamoumi
(1999)
Touririne Fm, Eastern
Anti-Atlas, Morocco
758S ‘frost dominated cold tidal
[surface]... display (sic) ice
wedges, desiccation polygons
and karstification’ Hamoumi
(1999)
Lower Touririne Sandstone member
contains trilobite fauna
comparable to ‘peltifer graptolite
Zone’ fauna of Letna
´Fm,
Bohemia; Upper Touririne
sandstone member coeval with
strata of clingani– ?complanatus
graptolite zone age; Touririne Fm
unconformably overlain by
Hirnantian tillites; stratigraphy in
Destombes et al. (1985)
(Continued)
BLACK SHALES IN THE EARLY PALAEOZOIC ICEHOUSE 141

Table 2. Continued
Stage Age Location Palaeolat. Evidence of ice Evidence for age
ASHGILL
(excluding
Hirnantian
sensu Cooper &
Sadler 2004)
‘Upper Ordovician’
Legrand (1985)
Tamadjert Fm*, Western
Hoggar, Algeria
608S buried landscapes show glacial
landforms, with striations,
drumlins, vestiges of
moraines, traces of
solifluction, ?pingos;
terrestrial and marine tillite
deposits (Beuf et al. 1971;
Biju-Duval et al. 1981;
Legrand 1985)
unconformably overlies strata of
Caradoc age (Biju-Duval et al.
1981); marine facies may contain
late Caradoc trilobites,
brachiopods and graptolites
(Gatinskiy et al. 1966); upper
part of formation contains
middle–late Llandovery
graptolites (Legrand 1970)
‘Upper Ordovician’
Tucker & Reid (1973)
Sierra Leone 508S ice-drop tillites with carbonate
boulders (Tucker & Reid
1973)
lithological similarity to and
correlation with similar deposits
in Guinea which are overlain by
Llandovery graptolite shales
(Tucker & Reid 1973)
Ashgill (Dore
´1981) Tillite de Feuguerolles,
Normandy, France
408S ice-drop tillites and diamictites
with glacially striated clasts
(Dore
´& Le Gall 1973; Dore
´
1981)
conformably underlain by strata
containing Caradoc age trilobites
and other fossils (Robardet et al.
1972; Dore
´& Le Gall 1973;
Dore
´1981); conformably
overlain by strata yielding latest
Ordovician– earliest Silurian
graptolite fauna (Phillipot &
Robardet 1971; Dore
´1981)
‘Late Ordovician’
Deynoux & Trompette
(1981)
Taoudeni Basin, West Africa 708S terrestrial and marine tillites in
the area near the Hodh with
striated boulders; glacially
reworked deposits with
outwash fans ‘similar to the
Icelandic sandur’ (Deynoux &
Trompette 1981);
‘micro-cordons’ that probably
represent subglacial eskers in
englacial tunnels; structures
similar to fentes minces
(Deynoux & Trompette 1981);
glacial pavements and roches
moutone
´es with striations,
furrows and crescentic
fractures in the Hodr;
glaciotectonic features
including ice-push ridges and
fractures en gradin
(Biju-Dival et al. 1974)
glacial deposits have erosive
disconformity at their base
(Deynoux & Trompette 1981;
Deynoux, Sougy & Trompette
1985) overlying the upper part of
Supergroup 2, which has an age
near the Cambro-Ordovician
boundary based on inarticulate
brachiopods and trace fossils
(Legrand 1969); base comparable
with Caradoc– Ashgill
unconformity in the Hoggar,
Tassilis & Anti-Atlas (Deynoux
& Trompette 1981); glacial
deposits conformably but
?diachronously overlain by
graptolite faunas of Upper
Ashgill–middle Llandovery age
(Underwood et al. 1998)
A. A. PAGE ET AL.142

HIRNANTIAN
sensu Cooper &
Sadler (2004)
extraordinarius graptolite
Zone acme (Sutcliffe
et al. 2000, 2001)
Northern Africa:
Upper 2nd Bani Fm,
Anti-Atlas Mts,
Morocco; Djebel Serraf
Fm, Ougarta Mts, Algeria
758S synchronous, large-scale tillite
and diamictite deposition; two
phases of regionally extensive
glaciomarine shelf sequences
and subglacial erosive
surfaces; ice-contact fans,
ice-rafted debris (Sutcliffe
et al. 2000, 2001; see also
Hamoumi 1999)
Hirnantia fauna within glaciogenic
deposits, and in underlying
formations (erosive contacts);
disconformably overlain by strata
containing Rhuddanian
graptolites (Destombes et al.
1985; Sutcliffe et al. 2000, 2001)
extraordinarius graptolite
Zone acme (Sutcliffe
et al. 2000, 2001)
Melez Chograne &
Memouniat Fms,
Libya
758S glaciomarine shelf deposits,
ice-rafted debris and erosive
surfaces covered by
ice-contact deposits (Havlı
´c
ˇek
& Massa 1973; Sutcliffe et al.
2001)
Hirnantia fauna throughout
succession, overlain by
Llandovery graptolites (Havlı
´c
ˇek
& Massa 1973; Sutcliffe et al.
2001)
extraordinarius graptolite
Zone acme (Sutcliffe
et al. 2000, 2001)
Tichit glacial group, the
Hodh, Mauritania
708S glacially striated dropstones,
diamictites (Deynoux &
Trompette 1981; Ghienne
2003)
dropstones coexist with graptolites
of Upper Ashgill age
(Underwood et al. 1998)
extraordinarius graptolite
Zone acme (Sutcliffe
et al. 2000, 2001)
Pakhuis Fm,
South Africa
158S tillites & diamictites, two
subglacial erosive surfaces
with striated pavements and
boulders, ice-rafted debris
(Rust 1981; Sutcliffe et al.
2000, 2001)
Hirnantia fauna in conformably
overlying Cedarberg Fm
(Sutcliffe et al. 2001)
?Hirnantian Tabuk Fm, Arabian
peninsula
508S diamictites & tillites with some
striated, faceted and polished
clasts; boulder pavements with
striations (McClure 1978)
tillite interfingers with late Caradoc
age graptolite shale also
containing trilobites of late
Caradoc or early Ashgill age
(Young 1981); overlying Qusaiba
Shale contains a Rhuddanian age
graptolite fauna (Lu
¨ning
et al. 2000)
Hirnantian (Armstrong
et al. 2005)
Ammar Fm, southern Jordan 508S tillite; glacial unconformity at
base and two episodes of
glacial incision;
conglomerates with glacially
faceted and striated clasts
(Abed et al. 1993)
conformably overlain by
persculptus graptolite Zone fauna
(Armstrong et al. 2005)
Hirnantian (Caputo 1998) Don Braulio Fm, Argentina 158S
‡
diamictite some striated, faceted
and polished pebbles and
cobbles (Bu
¨ggish & Astini
1993)
overlain by Hirnantia fauna
(Sutcliffe et al. 2001)
(Continued)
BLACK SHALES IN THE EARLY PALAEOZOIC ICEHOUSE 143

Table 2. Continued
Stage Age Location Palaeolat. Evidence of ice Evidence for age
?Hirnantian (Caputo
1998)
Iapo
´Fm, Parana
´Basin,
Brazil
258S diamictites with faceted and
striated clasts (Maack 1957;
Rocha-Campos 1981)
lithological comparison with South
African and Argentinian tillites
(Caputo 1998); Iapo
´Fm
discordantly overlies rhyolites of
the Castro Group dated at
450 +25 Ma (Bigarella 1970);
correlation with interfingering
Vila Maria Fm suggests
diamictite overlain by early
Llandovery palynomorph and
shelly fauna (Caputo & Crowell
1985)
Hirnantian Prague Basin, Czech
Republic
558S two intervals of diamictite
deposition (Brenchley &
S
ˇtorch 1989)
conformably underlain by
Mucronaspis fauna and anceps
Zone graptolites; conformably
overlain by Hirnantian fauna
(S
ˇtorch & Mergl 1989)
LLANDOVERY
(Rhuddanian)
?Rhuddanian San Gaba
´n–Cancan
˜iri–Zapla
Fms, Bolivia, Argentina &
Peru
558S widely extensive diamictites
with striated and faceted clasts
(Crowell et al. 1981; Caputo
& Crowell 1985;
Dı
´az-Martı
´nez & Grahn 2007)
oldest diamictite horizon overlain
by Aeronian chitinozoan fauna
and underlain by Rhuddanian
chitinozoan ; (Dı
´az-Martı
´nez &
Grahn 2007); these formations
unconformably overlie Caradoc
strata showing evidence of
Ashgillian deformation (Crowell
et al. 1981)
‘Upper Ordovician or
Lower Silurian’
(Kennedy 1981)
Stoneville & Beaver Cove
Fms, Newfoundland,
Canada
608S
†
diamictite beds; dropstones
probably derived by iceberg
rafting (Kennedy 1981)
Stoneville & Beaver Cove Fms are
coeval (Kennedy 1981); former
underlain by poorly preserved
corals of Upper Ordovician or
Lower Silurian age (McCann &
Kennedy 1974) and lithological
correlatives of its upper part have
yielded Llandovery age fossils
(Eastler 1969; Kennedy 1981)
LLANDOVERY
(Aeronian)
gregarius graptolite Zone
(Caputo 1998)
Nhamunda
´Fm, Amazonas
Basin Brazil
608S diamictite; ice-push & ice-shear
deformation structures
(Carozzi et al. 1973; Caputo
1998)
diamictite immediately overlain by
gregarius Zone graptolite fauna
and chitinozoan fauna (Grahn &
Paris 1992; Caputo 1998)
A. A. PAGE ET AL.144

‘late Aeronian– early
Telychian’ (Caputo
1998)
Nhamunda
´Fm, Amazonas
Basin, Brazil
608S diamictite; ice-push & ice-shear
deformation structures
(Carozzi et al. 1973; Caputo
1998)
overlies gregarius Zone fauna;
shales lateral to tillites yield an
early Telychian chitinozoan fauna
(Caputo 1998)
?Aeronian San Gaba
´n–Cancan
˜iri–Zapla
Fms, Bolivia, Argentina &
Peru
608S widely extensive diamictites
with striated and faceted clasts
(Crowell et al. 1981; Caputo
& Crowell 1985;
Dı
´az-Martı
´nez & Grahn 2007)
overlies shales yielding Aeronian
chitinozoans (Dı
´az-Martı
´nez &
Grahn 2007), overlain by shales
and a younger diamictite horizon
Llandovery (Caputo
1998)
Ipu Fm, Parnaı
´aba Basin &
Cariri Valley, Brazil
708S three diamictite layers (Caputo
& Crowell 1985; Grahn &
Caputo 1992); faceted pebbles
(Kegel 1953)
interfingers with Tiangua
´Fm,
which contains Early Silurian
chitinozoans and acritarchs
(Caputo & Lima 1984);
individual diamictites may
correlate with the better-dated
diamictites in the Nhamunda
´Fm
(Grahn & Caputo 1992)
LLANDOVERY
(Telychian,
including
centrifugus
Zone)
late Telychian (Grahn in
Cramer & Saltzman
2007)
Nhamunda
´Fm, Amazonas
Basin, Brazil
708S diamictites and tillites; ice-push
& ice-shear deformation
structures (Carozzi et al.
1973; Caputo 1998)
Late Telychian– early Wenlock
chitinozoan fauna in
interfingering shales (Caputo
1998); first appearance of
chitinozoa M. magaritana above
the youngest tillite (Grahn in
Cramer & Saltzman 2007)
Late Telychian;
(Dı
´az-Martı
´nez 2007)
San Gaba
´n–Cancan
˜iri–Zapla
Fms, Bolivia, Argentina &
Peru
658S widely extensive diamictites
with striated and faceted clasts
(Crowell et al. 1981; Caputo
& Crowell 1985;
Dı
´az-Martı
´nez & Grahn 2007)
youngest diamictite conformably
overlain by Sacla limestone,
which has early Wenlock age
based on occurrence of the
conodont O. sagitta rhenana
(Dı
´az-Martinez 2007); however,
acritarchs and chitinozoans in
intercalated shale horizons
suggest a Llandovery– Wenlock
boundary age (Sua
´rez-Soruco
1995); this diagnosis may suggest
an older age than Ireviken Datum
2 (see Appendix)
*Synonym of Felar Felar Fm.
†
This part of Nova Scotia is thought to have been on the margin of West Africa at this time (Kennedy 1981).
‡
Position of the Argentine Precordillera poorly constrained at this interval.
BLACK SHALES IN THE EARLY PALAEOZOIC ICEHOUSE 145

faunal turnover and changes in oceanic temperature and
oxygenation. As such, faunally derived sea-level curves
do not necessarily provide independent evidence of
glacioeustasy in this interval. We only make passing
reference to Loydell (1998), even though this offers well-
defined evidence of Silurian sea-level change with good
stratigraphic control. However, there is no curve defined
using a comparable method for the Ordovician. And, as
Loydell’s (1998) sea-level curve uses deposition of grap-
tolite shale as a criterion for establishing sea-level
change, employing it would preclude an independent test
of the relationship between sea-level and black shale
distribution during the EPI.
Though we have used sea-level curves for the
Ordovician and Silurian by different authors (Ross &
Ross 1996; Nielsen 2003a), both are compiled using the
same method and correlate sequences in North America
and Baltoscandia. As the Iapetus Ocean closed, there
may have been an increased local-tectonic component in
the Laurentian record of sea-level change during the EPI
(e.g. McKerrow et al. 2000). Nevertheless, correlation
between two palaeocontinents reduces the chance of con-
flating relative tectonic changes with global changes in
sea-level, and significant global events should register
above local noise. The Nielsen (2003a) sea-level curve
for the Ordovician shows a strong correlation with the
equivalent sea-level curve by Ross & Ross (1995), but
offers better stratigraphic resolution. Meanwhile, the
Ross & Ross (1996) curve for the Silurian employs a
method consistent with that used by Nielsen (2003a)in
the Ordovician. The original Ross & Ross (1996) sea-level
curve has poor biostratigraphic control in the late
Telychian (Loydell 1998), with the authors referring to
an undifferentiated crenulata Zone between the griesto-
niensis and centrifugus zones. This interval can be differ-
entiated into four graptolite zones (e.g. Loydell et al.
1998). As such, Figure 2 illustrates an amended version
of the Ross & Ross (1996) curve based on the recorrelation
of their original stratigraphy by Loydell (1998).
The sequence stratigraphic patterns observed by Ross
& Ross (1996) may be consistent with 20– 60 m changes
in sea-level in less than 1–2 Ma. The rates and frequency
of sea-level change during the EPI (Ross & Ross 1996;
Nielsen 2003a,b) are consistent with third-order sequence
stratigraphic cycles. They are therefore more likely
glacially than tectonically forced (Church & Coe 2003).
Oxygen isotopes: The oxygen isotope data presented in
Table 1 are obtained from the shells of brachiopods. These
were selected on the basis that they showed no trace
element or microstructural evidence of significant diage-
netic alteration (Brenchley et al. 1994; Marshall et al.
1997; Azmy et al. 1998; Tobin et al. 2005). Modern bra-
chiopods secrete a low-Mg calcite shell at or near isotopic
equilibrium with seawater (Lowenstam 1961; Carpenter &
Lohmann 1995; James et al. 1997), which tends to resist
diagenesis (Marshall et al. 1997; Azmy et al. 1998). The
isotopic composition of their shells shows little deviation
due to vital effects at the present day (Carpenter &
Lohmann 1995). Analysis of multi-taxa assemblages in
the Silurian suggests that vital effects may not have been
significant in the Palaeozoic (Samtleben et al. 2001).
Hence, these fossil data have been regarded as representa-
tive of the isotopic composition of Early Palaeozoic
seawater (e.g. Veizer et al. 1997; Samtleben et al. 2001).
Positive d
18
O excursions can be achieved by decreases
in temperature or increases in salinity (e.g. Hays &
Grossman 1991). The latter can be achieved due to
increased ice volume or decreased freshwater input, and
both may occur along with decreased temperature in gla-
ciations (e.g. Azmy et al. 1998; Tobin et al. 2005).
Though changes in salinity alone have been argued to
account for the positive isotope excursions observed in
this interval (e.g. Samtleben et al. 1996; Bickert et al.
1997), such changes represent salinity change of c.
14‰, which is an implausibly large change that cannot
be tolerated by brachiopods (Azmy et al. 1998). As
such, we interpret these positive excursions as represent-
ing decreases in temperature, which may be accompanied
by smaller increases in salinity. Therefore, positive d
18
O
excursions can be related to glaciations if consistent
with other evidence (e.g. Brenchley et al. 1994; Azmy
et al. 1998).
Carbon isotopes: The isotopic curves illustrated in
Figure 2 represent a consistent record of the d
13
C of car-
bonates throughout the EPI. These curves were compiled
using bulk rock carbonates in the Baltic region. These
strata have a well-resolved stratigraphy correlated to grap-
tolite, conodont and chitinozoan zones (e.g. Loydell et al.
1998, 2003; Loydell & Nestor 2005; No
˜lvak et al. 2006).
In this interval, the Estonian chronostratigraphic nomen-
clature is well-established across Baltoscandia and
widely used for correlation of carbonate-hosted assem-
blages (Cooper & Sadler 2004; Melchin et al. 2004).
This allows precise correlation between sea-level curves
and isotope data.
The d
13
C curves we illustrate were compiled by a
single research group using consistent sampling strategies
and analytic techniques on bulk rock carbonates (for
methods, see Kaljo et al. 1997, 1998, 2001). These data
show little evidence for diagenetic alteration (Kaljo
et al. 1997, 1998, 2001). The curves in Figure 2 are con-
sistent with each other. Individual excursions show good
accord with both organic and inorganic carbon-isotope
records derived in different regions (e.g. Underwood
et al. 1997; Kump et al. 1999; Melchin & Holmden
2006). The major positive excursions highlighted in
Table 1 and Figure 2 are recognized in the more-or-less
continuous record of carbon isotopes from the shells of
Silurian brachiopods (e.g. Azmy et al. 1998). They also
occur in the bulk rock data from the Ordovician of
Laurentia (e.g. Patzkowsky et al. 1997; Saltzman &
Young 2005) and the compilation for the Palaeozoic of
the Great Basin, USA (Saltzman 2005). The similarity of
the d
13
C curves in Figure 2 to these other d
13
C curves
highlights that the bulk-rock curves of Baltica record
global perturbations in the carbon cycle.
A. A. PAGE ET AL.146

Kump & Arthur (1999) and Kump et al. (1999) show
that positive carbon-isotope excursions can be achieved
by (a) increasing productivity, (b) increasing the burial
flux of organic carbon, or (c) by positive excursions in
the d
13
C value of riverine input into the marine carbon
reservoir due to increased terrestrial carbonate weathering.
These alternatives can be distinguished by analysis of
coupled patterns of organic and inorganic carbon isotopes
(cf. Kump & Arthur 1999; Kump et al. 1999). For
example, coupled organic and inorganic d
13
C data are
available for the Hirnantian glaciation, and these positive
d
13
C excursions have been interpreted as representing
changes in weathering (Kump et al. 1999; Melchin &
Holmden 2006). This excursion occurs during a major
regression (e.g. Brenchley et al. 2004; Melchin &
Holmden 2006), which exposed shallow marine carbon-
ates to terrestrial weathering, resulting in a more positive
d
13
C value of river waters (Kump et al. 1999). Given
that all the positive d
13
C excursions of the EPI correspond
to lowstands in cooler intervals with decreased organic
carbon burial in deep-water settings (Table 1; Fig. 2), all
positive d
13
C excursions may reasonably be interpreted
as being due to increased weathering of shallow carbon-
ates exposed in regressions.
The positive d
13
C excursion isotope excursion associ-
ated with the Hirnantian glaciation is hard to reconcile
with increased productivity or organic preservation (cf.
Brenchley et al. 1994). This event is coincident with a
mass extinction (e.g. Sutcliffe et al. 2000; Chen et al.
2005), when there is no evidence of increased organic
burial in even the deepest-water facies (e.g. Armstrong
& Coe 1997). We therefore believe it is best to consider
the positive carbon-isotope excursions of the EPI to rep-
resent cooler events if they are coincident with regression
(e.g. Patzkowsky et al. 1997; Azmy et al. 1998; Kaljo
et al. 2003; Tobin et al. 2005; Johnson 2006), and that
these excursions may be related to glaciations if consistent
with other evidence.
Oxic– anoxic stratigraphy: Anoxic intervals are
recognized from the occurrence of laminated hemipelagic
mudrocks in deep-water settings (i.e. below storm wave
base), which often contain graptolites. Graptolites are
organic walled macrozooplankton and the majority of
their fossil record comes from distal, anoxic mudrocks
(Chapman 1991; Underwood 1992; Finney & Berry
1997). They may also be sporadically found in oxic
facies, should they have undergone rapid burial. For
example, Stimulograptus sedgwickii occurs in well-
ventilated sandstones and siltstones in the Girvan area,
UK (Floyd & Williams 2003), and in the shelf successions
of the Llandovery area, Wales, UK (Cocks et al. 1984).
Likewise, tiny graptolite fragments have been reported
from rocks showing evidence of bioturbation (e.g. Arm-
strong & Coe 1997; Mullins & Aldridge 2004). Thus,
the presence of graptolites in well-laminated, dark-grey
or black mudrocks is evidence of anoxia rather than
graptolite palaeoecology (cf. Berry et al. 1987). Similarly,
the absence of macrofossil graptolites in poorly laminated
and/or burrowed paler grey shales is more typical of oxy-
genated bottom waters and sediments.
The oxic–anoxic stratigraphy for the EPI presented
here (Fig. 3 and refs therein) is derived from correlating
anoxic intervals in the deep-water record of UK succes-
sions. These successions are located in the Welsh Basin
and Southern Uplands of Scotland, which occur on the
Iapetus margins of Avalonia and Laurentia respectively
(Zalasiewicz 2001). They have a well-established,
high-resolution stratigraphy that allows such oxic–
anoxic transitions to be recognized at a sub-graptolite
zone resolution (e.g. Verniers & Vandenbroucke 2006).
Thus, if anoxic facies are deposited simultaneously in
the Welsh Basin and Southern Uplands, they represent
at least an Iapetus-wide anoxic event (Fig. 2). Where
intervals of anoxia in the Southern Uplands and Welsh
Basin do not correlate, it may be that sediment redox con-
ditions represent local rather than oceanic events. Leggett
(1980) made a similar compilation for the Early Palaeo-
zoic of the UK, employing stage level correlations.
However, our higher resolution oxic–anoxic stratigraphy
allows individual events to be correlated at a biozone
level (e.g. Fig. 3).
The basis of the UK record as a reliable record of
global marine anoxia requires consideration of the chan-
ging depositional settings of both the Welsh Basin and
Southern Uplands. We select data from intervals where
these strata were deposited in shelf and/or deep-basin
environments. To establish the global extent, we also
correlate this stratigraphy with redox changes recognized
in the deep-water record of the Rheic and/or
Palaeothethys Oceans. As far as we are aware, the
record of marine anoxia in the deep-water facies of the
UK represents the only well-dated, continuous succession
where the oxic–anoxic stratigraphy has been sufficiently
documented to assemble a composite oxic–anoxic strati-
graphy for the EPI.
The Welsh Basin was a restricted basin on the eastern
margin of Avalonia. Its depositional setting and sedimen-
tary history are reviewed by Woodcock (2000) and
Zalasiewicz (2001). Local changes in freshwater run-off,
nutrient input or upwelling may have induced localized
anoxic events by altering productivity and stratification.
Its sedimentary record stretches from the Cambrian to
latest Silurian. Nevertheless, the widespread volcanism
prior to the mid-Caradoc significantly disrupted patterns
of marine topography, subsidence and deposition
(Woodcock 1990). This resulted in a more ambiguous
and locally variable pattern of basin redox conditions at
this time (cf. Leggett 1980). Once sediment input outpaced
subsidence in the late Silurian (King 1994; Woodcock
2000), the basin became increasingly shallow and it rapidly
filled with sediment. The oxic –anoxic stratigraphy of the
Welsh Basin correlates extremely well with similar strati-
graphies where available for the deep-water successions of
the Howgill Fells and Lake District, Northern England
(cf. Rickards 1970; Hutt 1974; Rickards & Woodcock
2005). Thus, the Welsh Basin provides a well-resolved
BLACK SHALES IN THE EARLY PALAEOZOIC ICEHOUSE 147

record of the oxic–anoxic stratigraphy of eastern Avalonia
throughout the EPI.
In contrast, the Moffat Shale Group of the Southern
Uplands is commonly thought to represent an ocean
floor environment approaching a trench (Leggett et al.
1979; Leggett 1987). It may have been susceptible to
changes in upwelling, which could have induced
localized anoxia by increasing export production (e.g.
Finney & Berry 1997). The Southern Uplands record of
the EPI is contained within an accretionary prism
formed from Iapetus thrust-slices during the Caledonian
Orogeny (Leggett et al. 1979; Leggett 1987; Strachan
2000). The succession has an early Caradoc to late Llan-
dovery age (Leggett 1980, 1987; Strachan 2000). Sub-
sequently, the Moffat Shale Group was overlain by the
massive flysch deposits of the Gala and Hawick groups
of late Llandovery to Wenlock age (White et al. 1991).
At a larger scale, the mudstones of the Moffat Shale
Group comprise a generally distal, condensed succession
of Ordovician and early Llandovery age, with slightly
more expanded and proximal facies in the mid- and late
Llandovery. So, rather than being a restricted basin, the
Southern Uplands provide a well-resolved record of
open marine conditions throughout all but the terminal
part of the EPI.
The major anoxic events of the Welsh Basin correlate
well with those from the Southern Uplands where data are
available (Fig. 3). Hence, there is no reason to believe that
the late Llandovery–Wenlock of the Welsh Basin is unre-
presentative of the Iapetus Ocean redox conditions,
especially as synchronous changes are seen in northern
England (cf. Rickards 1970; Rickards & Woodcock
2005). Thus, the oxic– anoxic stratigraphies of the UK
basins (Fig. 3), from which we compiled the summary
oxic– anoxic stratigraphy (Fig. 2), probably represent the
best record available of oxic–anoxic transitions in low-
latitude deep waters during the EPI.
References
ABED, A. M., MAKHLOUF, I. M., A MIREH,B.S.&
KHALIL, B. 1993. Upper Ordovician glacial deposits
in southern Jordan. Episodes,16, 316–328.
AINSAAR, L., MEIDLA,T.&MARTMA, T. 1999. Evi-
dence for a widespread carbon isotopic event associ-
ated with late Middle Ordovician sedimentological
and faunal changes in Estonia. Geological Magazine,
136, 49– 62.
ALDRIDGE, R. J., JEPPSSON,L.&DORNING, K. J. 1993.
Early Silurian oceanic episodes and events. Journal of
the Geological Society, London,150, 501– 513.
AOUDEH,S.M.&AL-HAJRI, S. A. 1995. Regional distri-
bution and chronostratigraphy of the Qusaiba Member
of the Qalibah Formation in the Nafud Basin, north-
western Saudi Arabia. In:A
L-HUSSEINI, M. I. (ed.)
Geo ’94, The Middle East Petroleum Geosciences,
vol. 1. Gulf PetroLink, Manama, Bahrain, 143 –154.
ARMSTRONG,H.A.&COE, A. L. 1997. Deep sea sedi-
ments record the geophysiology of the end Ordovician
glaciation. Journal of the Geological Society, London,
154, 929– 934.
ARMSTRONG, H. A., TURNER, B. R., MAKHLOUF, I. M.,
WEEDON, G. P., WILLIAMS, M., ALSMADI,A.&
ABU SALAH, A. 2005. Origin, sequence stratigraphy
and depositional environment of an Upper Ordovician
(Hirnantian) deglacial black shale, Jordan. Palaeogeo-
graphy, Palaeoclimatology, Palaeoecology,220,
273–289.
ARMSTRONG, H. A., TURNER, B. R., MAKHLOUF, I. M.,
WEEDON, G. P., WILLIAMS, M., ALSMADI,A.&
ABU SALAH, A. 2006. Reply to ‘Origin, sequence
stratigraphy and depositional environment of an
upper Ordovician (Hirnantian) deglacial black shale,
Jordan’. Palaeogeography, Palaeoclimatology,
Palaeoecology,230, 356– 360.
ARRHENIUS, S. 1896. On the influence of carbonic acid in
the air upon the temperature on the ground. Philoso-
phical Magazine and Journal of Science,41, 237 –275.
AZMY, K., VEIZER, J., BASSETT,M.G.&COPPER,P.
1998. Oxygen and carbon isotopic composition of
Silurian brachiopods: Implications for coeval seawater
and glaciations. Geological Society of America Bulle-
tin,110, 1499– 1512.
BAKER, S. J. 1981. The graptolite biostratigraphy of a
Llandovery outlier near Llanystumdwy, Gwynedd,
North Wales. Geological Magazine,118, 355– 365.
BARKER, S., HIGGINS,J.A.&ELDERFIELD, H. 2003.
The future of the carbon cycle: review, calcification
response, ballast and feedback on atmospheric CO
2
.
Philosophical Transactions of the Royal Society of
London, series A,361, 1977– 1999.
BARNES,C.R.&BOLTON, T. E. 1998. The Ordovician–
Silurian boundary on Manitoulin Island, Ontario,
Canada. Bulletin of the British Museum, Natural
History, Geology,43, 247– 253.
BERNER, R. A. 1991. A model for atmospheric CO
2
over
Phanerozoic time. American Journal of Science,291,
339–376.
BERNER, R. A. 1994. GEOCARB II: A revised model of
atmospheric CO
2
over Phanerozoic time. American
Journal of Science,294, 56–91.
BERNER, R. A. 1998. 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.
BERNER, R. A. 2001. Modelling atmospheric O
2
over
Phanerozoic time. Geochimica et Cosmochimica
Acta,65, 685– 694.
BERNER, R. A. 2003. The long-term carbon cycle, fossil
fuels and atmospheric composition. Nature,426,
323–326.
BERNER, R. A. 2006. GEOCARBSULF: A combined
model for Phanerozoic atmospheric O
2
and CO
2
over
Phanerozoic time. Geochimica et Cosmochimica
Acta,70, 5653– 5664.
BERNER,R.A.&CANFIELD, D. E. 1989. A new model
for atmospheric oxygen over Phanerozoic time.
American Journal of Science,289, 333– 361.
BERNER,R.A.&KOTHAVALA, Z. 2001. Geocarb III: A
revised model of atmospheric CO
2
over Phanerozoic
time. American Journal of Science,301, 182– 204.
BERNER,R.A.&RAISWELL, R. 1983. Burial of organic
carbon and pyrite sulfur in sediments over Phanerozoic
A. A. PAGE ET AL.148

time: a new theory. Geochimica et Cosmochimica
Acta,47, 855– 862.
BERNER, R. A., LASAGA,A.C.&GARRELS, R. M. 1983.
The carbonate–silicate geochemical cycle and its
effect on atmospheric carbon dioxide over the past
100 million years. American Journal of Science,283,
641–683.
BERRY, W. B. N., WILDE,P.&QUINBY-HUNT,M.S.
1987. The oceanic non-sulfidic oxygen minimum
zone; a habitat for graptolites? Bulletin of the Geologi-
cal Society of Denmark (Meddelelser fra Dansk
Geologisk Forening),35, 103– 114.
BEUF, S., BIJU-DUVAL, B., DE CHARPA L, O., ROGNON,
P., GARIEL,O.&BENNACEF, A. 1971. Les gre
`sdu
Pale
´ozoı
¨que infe
´rieur au Sahara (Se
´dimentation et
discontinuities, evolution structurale d’un craton).
Publication de l’Institut Franc¸ais Pe
´trole, 18.
BICKERT, T., PA
¨TZOLD, J., SAMTLEBEN,C.&
MUNNECKE, A. 1997. Paleoenvironmental changes
in the Silurian indicated by stable isotopes in brachio-
pod shells from Gotland, Sweden. Geochimica et
Cosmochimica Acta,61, 2717– 2730.
BIGARELLA, J. J. 1970. Continental drift and palaeo-
current analysis. Second Gondwana Symposium,
Proceedings and Papers. Council for Scientific and
Industrial Research, South Africa, 73–97.
BIJU-DUVAL, B., DEYNOUX,M.&ROGNON, P. 1974.
Essai d’interpre
´tation de ‘fractures en gradins’ obser-
ve
´es dans les formations glaciaires Pre
´cambriennes
et Ordoviciennes du Sahara. Revue de Geographie
Physique et de Geologie Dynamique,26, 503– 512.
BIJU-DIVAL, B., DEYNOUX,M.&ROGNON, P. 1981.
Late Ordovician tillites of the Central Sahara. In:
HAMBREY,M.J.&HARLAND, W. B. (eds) Earth’s
Pre-Pleistocene Glacial Record. Cambridge Univer-
sity Press, Cambridge, 99–107.
BJERRESKOV, M. 1975. Llandoverian and Wenlockian
graptolites from Bornholm. Fossils and Strata,8,
1–94.
BRENCHLEY, P. J. 1988. Environmental changes close to
the Ordovician– Silurian boundary. Bulletin of the
British Museum of Natural History (Geology), 43,
377–385.
BRENCHLEY,P.J.&S
ˇTORCH, P. 1989. Environmental
changes in the Hirnantian (upper Ordovician) of the
Prague Basin, Czechoslovakia. Geological Journal,
24, 165– 181.
BRENCHLEY, P. J., MARSHALL,J.D.&CARD EN,
G. A. F. ET AL. 1994. Bathymetric and isotopic evi-
dence for a short-lived Late Ordovician glaciation in
a greenhouse period. Geology,22, 295– 298.
BRENCHLEY, P. J., CARDEN, G. A., HINTS,L.ET AL.
2003. High-resolution isotope stratigraphy of Late
Ordovician sequences: constraints on the timing of
bio-events and environmental changes associated
with mass extinction and glaciation. Geological
Society of America, Bulletin,115, 89– 104.
BROECKER, W. S. 1982. Glacial to interglacial changes in
ocean chemistry. Progress in Oceanography,11,
151–197.
BRUCKSCHEN, P., OESMANN,S.&VEIZER, J. 1999.
Isotope stratigraphy of the European Carboniferous:
proxy signals for ocean chemistry, climate and tec-
tonics. Chemical Geology,161, 127– 163.
BRUNTON, F. R., SMITH, L., D IXON, O. A., COPPER, P.,
NESTOR,H.&KERSHAW, S. 1998. Silurian reef epi-
sodes, changing seascapes, and paleobiogeography. In:
LANDING,E.&JOHNSON, M. E. (eds) Silurian
Cycles, Linkages of Dynamic Stratigraphy with Atmos-
pheric, Oceanic, and Tectonic Changes. New York
State Museum Bulletin, 491, 265–282.
BU
¨GGISH,W.&ASTINI, R. 1993. The Late Ordovician
ice age: new evidence from the Argentine Precordil-
lera. In:F
INDLAY, R. H., UNRUG, R., BANKS,M.
R. & VEEVERS, J. J. (eds) Gondwana Eight – Assem-
bly, Evolution and Dispersal. Balkema, Rotterdam,
439–447.
BUTTERFIELD, N. J. 1997. Plankton ecology and the Pro-
terozoic–Phanerozoic transition. Paleobiology,23,
247–262.
CALNER,M.&ERIKSSON, M. E. 2006. Silurian research
at the crossroads. GFF,128, 73–74.
CALNER, M., JEPPSSON,L.&MUNNECKE, A. 2004. The
Silurian of Gotland – Part I: Review of the strati-
graphic framework, event stratigraphy, and
stable carbon and oxygen isotope development.
Erlanger Geologische Abhandlungen, Sonderband,5,
113–131.
CAPUTO, M. V. 1998. Ordovician –Silurian glaciations
and global sea-level changes. In:L
ANDING,E.&
JOHNSON, M. E. (eds) Silurian Cycles, Linkages of
Dynamic Stratigraphy with Atmospheric, Oceanic,
and Tectonic Changes. New York State Museum Bul-
letin, 491, 15– 25.
CAPUTO,M.V.&CROW ELL, J. C. 1985. Migration of
glacial centers across Gondwana during Paleozoic
Era. Geological Society of America, Bulletin,96,
1020– 1036.
CAPUTO,M.V.&LIMA, E. C. 1984. Estratigrafia, idade e
correlac¸a
˜o do Grupo Serra Grande – Bacia do Par-
naı
´ba. Congresso Brasileiro de Geologia,33,
740–753.
CAROZZI, A. V., PAMPLONA, H. R. P., C ASTRO,J.C.&
CONTREIRAS, C. J. A. 1973. Ambientes deposicionais
e evoluc¸a
˜o tecto-sedimentar da sec¸a
˜o ela
´stica paleo-
zo
´ica da Bacia do Me
´dio Amazonas. Congresso Brasi-
leiro de Geologia,27, 279– 314.
CARPENTER,S.J.&LOHMANN, K. C. 1995. d
18
O and
d
13
C values of modern brachiopods. Geochimica et
Cosmochimica Acta,59, 3749– 3764.
CATLING,D.C.&CLAIRE, M. 2005. How Earth’s atmos-
phere evolved to an oxic state: A status report. Earth
and Planetary Science Letters,237, 1–20.
CAVE, R. 1979. Sedimentary environments of the basinal
Llandovery of mid-Wales. In:H
ARRIS, A. L.,
HOLLAND,C.H.&LEAKE, B. E. (eds) The Caledo-
nides of the British Isles – Reviewed. Special Publi-
cation of the Geological Society, London, 8, 517–26.
CAVE,R.&DIXON, R. J. 1993. The Ordovician and Silur-
ian of the Welshpool area. In:W
OODCOCK,N.H.&
BASSETT, M. G. (eds) Geological Excursions in
Powys, Central Wales. Geological Series – National
Museum of Wales, 14,51–84
CHAMBERLIN, T. C. 1899. An attempt to frame a working
hypothesis of the cause of glacial periods on an atmos-
pheric basis. Journal of Geology,7, 545–584.
CHAPMAN, A. J. 1991. How are they preserved? In:
PALMER,D.&RICKARDS, R. B. (eds) Graptolites:
BLACK SHALES IN THE EARLY PALAEOZOIC ICEHOUSE 149

Writing in the Rocks. Fossils illustrated, vol. 1. The
Boydell Press, Woodbridge, UK, 6– 11.
CHEN X., RONG JIAYU,C.E.MITCHELL, D.A.T.
HARPER,FAN,JUNXUAN,ZHAN,RENBIN,ZHA NG,
YUANDONG,LI,RONGYU &WANG,YI2000. Late
Ordovician to earliest Silurian graptolite and brachio-
pod biozonation from the Yangtze region, South
China, with a global correlation. Geological Maga-
zine,137, 623– 650.
CHEN, X., MELCHIN, M. J., SHEETS, H. D., MITCHELL,
C. E. & FAN, J.-X. 2005. Patterns and Processes of
latest Ordovician graptolite extinction and recovery
based on data from South China. Journal of Paleontol-
ogy,79, 842– 861.
CHURCH,K.D.&COE, A. L. 2003. Processes controlling
relative sea-level change and sediment supply. In:
COE, A. L. (ed.) The Sedimentary Record of Sea-Level
Change. Cambridge University Press, Cambridge,
99–117.
COCKS,L.R.M.&TORSVIK, T. H. 2002. Earth geogra-
phy from 500 to 400 million years ago: A faunal and
palaeomagnetic review. Journal of the Geological
Society of London,159, 631– 644.
COCKS, L. R. M., WOODCOCK, N. H., RICKARDS, R. B.,
TEMPLE,J.T.&LANE, P. D. 1984. The Llandovery
Series of the type area. Bulletin of the British
Museum (Natural History): Geology,38, 131–182.
COHEN, A. S., COE, A. L., HARDING,S.M.&
SCHWARK, L. 2004. Osmium isotope evidence for
the regulation of atmospheric CO
2
by continental
weathering. Geology,32, 157– 160.
COOPER,R.A.&SADLER, P. M. 2004. The Ordovician
Period. In:G
RADSTEIN, F. M., OGG,J.G.&
SMITH, A. G. (eds) A Geologic Time Scale 2004.
Cambridge University Press, Cambridge, 165–187.
COPPER, P. 1994. Ancient reef ecosystem expansion and
collapse. Coral Reefs,13, 3–11.
CRAMER,B.D.&SALTZMAN, M. R. 2005. Sequestration
of
12
C in the deep ocean during the early Wenlock
(Silurian) positive carbon isotope excursion. Palaeo-
geography, Palaeoclimatology, Palaeoecology,219,
333–349.
CRAMER,B.D.&SALTZMAN, M. R. 2007. Fluctuations
in epeiric sea carbonate production during
Silurian positive carbon isotope excursions: A review
of proposed paleoceanographic models: Palaeo-
geography, Palaeoclimatology, Palaeoecology,
in press.
CROWELL, J. C., SUA
´REZ-SORUCO,R.&ROCHA-
CAMPOS, A. C. 1981. The Silurian Cancan
˜iri (Zapla)
Formation of Bolivia, Argentina and Peru. In:
HAMBREY,M.J.&HARLAND, W. B. (eds) Earth’s
Pre-Pleistocene Glacial Record. Cambridge Univer-
sity Press, Cambridge, 902– 907.
CROWLEY,T.J.&BERNER, R. A. 2001. CO
2
and climate
change. Science,292, 870– 872.
DAVIES,J.R.&WATER S, R. A. 1995. The Caban
Conglomerate and Ystrad Meurig Grits Formation –
nested channels and lobe switching on a mud domi-
nated latest Ashgill to Llandovery lope-apron, Welsh
Basin, UK. In:P
ICKERING, K. T., HISCOTT, R. N.,
KENYON, N. H., RICCI LUCCHI,F.&SMITH,
R. D. A. (eds) Atlas of Deep Water Environments:
Architectural Style in Turbidite Systems. London:
Chapman & Hall, 182–93.
DAVIES, J. R., FLETCHER, C. J. N., WAT ERS, R. A.,
WILSON, D., WOODHALL,D.G.&ZALASIE WICZ,
J. A. 1997. Geology of the country around Llanilar
and Rhayader. Memoir of the British Geological
Survey, Sheets 178 and 179 (England and Wales).
HMSO, London.
DAVIES,J.R.,WATERS,R.A.,WILBY,P.R.,WILLIAMS,
M. & WILSON, D. 2003. Geology of Cardigan and
Dinas Island district – a brief explanation of the geologi-
cal map. Sheet explanation of the British Geological
Survey. 1: 50000 sheet 193 (including part of sheet
210) Cardigan & Dinas Island (England & Wales).
Keyworth, Nottingham, British Geological Survey.
DESTOMBES, J., HOLL AND,H.&WILLEFERT, S. 1985.
Lower Palaeozoic rocks of Morocco. In:H
OLLAND,
C. H. (ed.) Lower Palaeozoic of north-western and
west-central Africa. Lower Palaeozoic Rocks of the
World, 4, 91–337, John Wiley & Sons, Chichester.
DEYNOUX,M.&TROMPETTE, R. 1981. Late Ordovician
tillites of the Taodeni Basin, West Africa. In:
HAMBREY,M.J.&HARLAND, W. B. (eds) Earth’s
pre-Pleistocene Glacial Record. Cambridge Univer-
sity Press, 89–96.
DEYNOUX, M., SOUGY,J.&TROMPETTE, R. 1985.
Lower Palaeozoic rocks of west Africa and the
western part of central Africa. In:H
OLLAND,C.H.
(ed.) Lower Palaeozoic of North-Western and West-
Central Africa. Lower Palaeozoic Rocks of the
World, 4, 337–497, John Wiley & Sons, Chichester.
DIAZ-MARTINEZ, E. 2007. The Sacta Limestone Member
(early Wenlock): cool-water, temperate carbonate
deposition at the distal foreland of Gondwana’s
active margin, Bolivia. Palaeogeography, Palaeocli-
matology, Palaeoecology, in press.
DIAZ-MARTINEZ,E.&GRAHN, Y. 2007. Early Silurian
glaciation along the western margin of Gondwana
(Peru, Bolivia, and northern Argentina): Palaeogeo-
graphic and geodynamic setting. Palaeogeography,
Palaeoclimatology, Palaeoecology, in press.
DORE
´,F.&LEGALL, J. 1973. Presence et position stra-
tigraphique de la tillite Ordovicienne dans le Maine (E.
du Massif Armoricain). Bulletin de la Socie
´te
´Ge
´ologi-
que de France,15, 32–33.
DORE
´, F. 1981. The late Ordovician tillite in Normandy
(Armorican Massif) In:H
AMBREY,M.J.&
HARLAND, W. B. (eds) Earth’s Pre-Pleistocene
Glacial Record. Cambridge University Press,
Cambridge, 582– 584.
EASTLER, T. E. 1969. Silurian geology of the Change
Islands and eastern Notre Dame Bay, Newfoundland.
Memoir of the American Association of Petroleum
Geologists,12, 425– 432.
ELDERFIELD, H. 2002. Climate change: Carbonate mys-
teries. Science,296, 1618– 1621.
EVANS, D. A. D. 2003. A fundamental Precambrian–
Phanerozoic shift in Earth’s glacial style? Tectonophy-
sics,375, 353– 385.
EYLES, N. 1993. Earth’s glacial record and its tectonic
setting. Earth Science Reviews,35, 1–248.
FINNEY,S.C.&BERRY, W. B. N. 1997. New perspec-
tives on graptolite distributions and their use as indi-
cators of platform margin dynamics. Geology,25,
919–922.
FLOYD,J.D.&WILLIAMS, M. 2003 (for 2002). A revised
correlation of Silurian rocks in the Girvan district, SW
A. A. PAGE ET AL.150

Scotland. Transactions of the Royal Society of Edin-
burgh, Earth Sciences,25, 383– 392.
FORTEY,R.A.&COCKS, L. R. M. 2003. Palaeontologi-
cal evidence bearing on global Ordovician–Silurian
continental reconstructions. Earth Science Reviews,
61, 245– 307.
FORTEY,R.A.&COCKS, L. R. M. 2005. Late Ordovician
global warming – The Boda Event. Geology,33,
405–408.
FORTEY, R. A., HARPER, D. A. T., INGHAM, J. K.,
OWEN,A.W.&RUSHTON, A. W. A. 1995. A revision
of Ordovician series and stages from the historical type
area. Geological Magazine,132, 15– 30.
FORTEY, R. A., HARPER, D. A. T., INGHAM, J. K.,
OWEN, A. W., PARKES, M. A., RUSHTON,A.W.A.
&W
OODCOCK, N. H. 2000. A revised correlation of
Ordovician rocks in the British Isles. Geological
Society of London Special Report,24, 1–83.
FRAKES, L. A., FRANCIS,J.E.&SYKT US, J. I. 1992.
Climate Modes of the Phanerozoic. Cambridge Uni-
versity Press, Cambridge.
GATINSKIY, Y. G., KLOCHKO, V. P., R OZMAN,K.S.&
TROFIMOV, D. M. 1966. Novyye dannyye po strati-
grafii paleozoyskikh otlozheniy yuzhnoy Sakhary.
Akadwmii Nauk SSSR, Doklady,170, 1154– 1157.
GENSEL,P.G.&EDWARDS, D. 2001. Plants Invade the
Land: Evolutionary and Environmental Perspectives.
Columbia University Press, New York.
GHIENNE, J.-F. 2003. Late Ordovician sedimentary
environments, glacial cycles, and post-glacial trans-
gression in the Taoudeni Basin, West Africa. Palaeo-
geography, Palaeoclimatology, Palaeoecology,189,
117–145.
GIBBS, M. T., BICE, K. L., BARRON,E.J.&KUMP,L.R.
2000. Glaciation in the early Paleozoic ‘greenhouse’;
the roles of paleogeography and atmospheric CO
2
.
In:H
UBER, B. T., MACLEOD,K.G.&WING,S.L.
(eds) Warm Climates in Earth History. Cambridge
University Press, Cambridge, 386– 422.
GOLDMAN, D., LESLIE, S. A., N O
˜LVAK,J.&YOUNG,S.
2005. The Black Knob Ridge section, southeastern
Oklahoma, USA: a possible Global Stratotype-Section
and Point (GSSP) for the base of the Middle Stage of
the Upper Ordovician Series. www.stratigraphy.org/
BKR.pdf.
GRAHN,Y.&CAPUTO, M. V. 1992. Early Silurian glacia-
tions in Brazil. Palaeogeography, Palaeoclimatology,
Palaeoecology,99, 9– 15.
GRAHN,Y.&PARIS, F. 1992. Age and correlation of the
Trombetas Group, Amazonas Basin, Brazil. Revue de
Micropale
´ontologie,35, 197– 209.
GUTIE
´RREZ-MARCO,J.C.&S
ˇTORCH, P. 1998. Grapto-
lite biostratigraphy of the Lower Silurian (Llandovery)
shelf deposits of the Western Iberian Cordillera, Spain.
Geological Magazine,135, 71– 92.
HAMOUMI, N. 1999. Upper Ordovician glaciation spread-
ing and its sedimentary record in Moroccan North
Gondwana margin. In:K
RAFT,P.&FJA
¨CKA,O.
(eds) Quo Vadis Ordovician? Acta Universitatis
Carolinae, Geologica, 43, 111– 114.
HATCH, J. R., JACOBSON, S. R., WITZ KE,B.J. ET AL.
1987. Possible late Middle Ordovician organic
carbon isotope excursion: evidence from
Ordovician oils and hydrocarbon source rocks, mid-
continent and east-central United States. American
Association of Petroleum Geologists, Bulletin,71,
1342– 1354.
HAVLI
´C
ˇEK,V.&MASSA, D. 1973. Brachiopodes de
l’Ordovicien Supe
´rieur de Libye occidental: impli-
cations stratigraphiques regionales. Geobios,6,
267–290.
HAYES, J. M., STRAUSS,H.&KAUFMAN, A. J. 1999.
The abundance of
13
C in marine organic matter and
isotopic fractionation in the global biogeochemical
cycle of carbon during the past 800 Ma. Chemical
Geology,161, 103– 125.
HAYS,P.D.&GROSSMAN, E. L. 1991. Oxygen isotopes
in meteoric calcite cements as indicators of continental
paleoclimate. Geology,19, 441– 444.
HAYWOOD, A. M., DEKENS, P., R AVELO,A.C.&
WILLIAMS, M. 2005. Warmer tropics during the
mid-Pliocene? Evidence from alkenone paleo-
thermometry and a fully coupled ocean-atmosphere
GCM. Geochemistry, Geophysics, Geosystems,6,
Q03010.
HERRMANN, A. D., PATZKOWSKY,M.E.&POLLARD,
D. 2003. Obliquity forcing with 8 –12 times preindus-
trial levels of atmospheric pCO
2
atmospheric CO
2
during the Late Ordovician glaciation. Geology,31,
485–488.
HERRMANN, A. D., PATZKOWSKY,M.E.&POLLARD,
D. 2004a. The impact of paleogeography, pCO
2
, pole-
ward ocean heat transport and sea level change on
global cooling during the Late Ordovician. Palaeogeo-
graphy, Palaeoclimatology, Palaeoecology,206,
59–74.
HERRMANN, A. D., HAUPT, B. J., P ATZKOWSKY, M. E.,
SEIDOV,D.&SLINGERLAND, R. L. 2004b. Response
of Late Ordovician paleoceanography to changes in
sea level, continental drift, and atmospheric pCO
2
:
potential causes for long-term cooling and glaciation.
Palaeogeography, Palaeoclimatology, Palaeoecol-
ogy,210, 385– 401.
HOLLAND, H. D. 1978. The Chemistry of the Atmosphere
and the Ocean. Wiley Interscience, New York.
HOVIUS, N. 1998. Controls on sediment supply by large
rivers. In:S
HANLEY,K.W.&MCCABE, P. J. (eds)
Relative Role of Eustasy, Climate, and Tectonism in
Continental Rocks. SEPM Special Publication, 59,
3–16.
HUTT, J. E. 1974. The Llandovery graptolites of the
English Lake District. Part 1. Monograph of the
Palaeontographical Society, London,128, 1– 56.
JAMES, N. P., BONE,Y.&KYSER, T. K. 1997. Brachio-
pod d
18
O values do reflect ambient oceanography:
Lacepede Shelf, southern Australia. Geology,25,
551–554.
JEPPSSON, L. 1990. An oceanic model for lithological and
faunal changes tested on the Silurian record. Journal of
the Geological Society, London,147, 663– 674.
JEPPSSON, L. 1997. The anatomy of the Mid–Early Silur-
ian Ireviken Event and a scenario for P – S events. In:
BRETT,C.E.&BAIRD, G. C. (eds) Paleontological
Events: Stratigraphic, Ecological, and Evolutionary
Implications. Columbia University Press, New York,
451–492.
JEPPSSON, L., ALDRIDGE,R.J.&DORNING, K. J. 1995.
Wenlock (Silurian) oceanic episodes and events.
BLACK SHALES IN THE EARLY PALAEOZOIC ICEHOUSE 151

Journal of the Geological Society, London,152,
487–498.
JOHNSON, M. E. 1996. Stable cratonic sequences and a
standard for Silurian eustasy. In:W
ITZKE, B. J., LUD-
VIGSON,G.A.&DAY, J. (eds) Paleozoic Sequence
Stratigraphy: Views from the North American
Craton. Geological Society of America Special
Paper, 306, 203– 211.
JOHNSON, M. E. 2006. Relationship of Silurian sea-level
fluctuations to oceanic episodes and events. GFF,
128, 115– 122.
JOHNSON, M. E., KALJO,D.L.&RONG, J.-Y. 1991.
Silurian eustasy. In:B
ASSETT, M. G., LANE,P.D.
&E
DWARDS, D. (eds) The Murchison Symposium.
Special Papers in Palaeontology, 44, 145–163.
KALJO,D.&MARTMA, T. 2006. Application of carbon
isotope stratigraphy to dating the Baltic Silurian
rocks, GFF,128, 123– 129.
KALJO, D., MARTMA, T., MANNIK,P.&VIIRA, V. 2003.
Implications of Gondwana glaciations in the Baltic
Late Ordovician and Silurian and a carbon isotopic
test of environmental cyclicity. Bulletin de la Socie
´te
´
Ge
´ologique de France,174, 59– 66.
KALJO, D., HINTS, L., H INTS, O., MARTMA,T.&
NO
˜LVAK, J. 1999. Carbon isotope excursions and
coeval biotic-environmental changes in the late
Caradoc and Ashgill of Estonia. In:K
RAFT,P.&
FJAKA, O. (eds) Quo Vadis Ordovician? Acta Univer-
sitatis Carolinae, Geologica, 43, 507–510.
KALJO, D., HINTS, L., MARTMA,T.&NO
˜LVAK, J. 2001.
Carbon isotope stratigraphy in the latest Ordovician of
Estonia. Chemical Geology,175, 49– 59.
KALJO, D., HINTS, L., MARTMA, T., NO
˜LVAK,J.&
ORASPO
˜LD, A. 2004. Late Ordovician carbon
isotope trend in Estonia, its significance in stratigraphy
and environmental analysis. Palaeogeography,
Palaeoclimatology, Palaeoecology,210, 165– 185.
KALJO, D., KIIPLI,T.&MARTMA, T. 1997. Carbon
isotope event markers through the Wenlock –Pridoli
sequence in Ohesaare (Estonia) and Priekule
(Latvia). Palaeogeography, Palaeoclimatology,
Palaeoecology,132, 211– 224.
KALJO, D., KIIPLI,T.&MARTMA, T. 1998. Correlation
of carbon isotope events and environmental cyclicity
in the East Baltic Silurian. In:L
ANDING,E.&
JOHNSON, M. E. (eds) Silurian Cycles: Linkages of
Dynamic Stratigraphy with Atmospheric, Oceanic
and Tectonic Changes. New York State Museum Bul-
letin, 491, 297– 312.
KALJO, D., MARTMA,T.&SAADRE, T. 2007. Post-
Hunnebergian Ordovician carbon isotope trend in Bal-
toscandia, its environmental implications and some
similarities with that of Nevada. Palaeogeography,
Palaeoclimatology, Palaeoecology, in press.
KEGEL, W. 1953. Contribuic¸a
˜o para o estudo do. Devo-
niano da Bacia do Parnaı
´ba. Boletim da Divisa
˜ode
Geologia e Mineralogia, Rio de Janeiro,141, 1–48.
KENNEDY, M. J. 1981. The early Palaeozoic Stoneville
Formation, northern Newfoundland. In: HAMBREY,
M. J. & HARLAND, W. B. (eds) Earth’s pre-
Pleistocene Glacial Record. Cambridge University
Press, Cambridge, 713– 716.
KING, L. M. 1994. Subsidence analysis of eastern Avalo-
nian sequences; implications for Iapetus closure.
Journal of the Geological Society, London,151,
647–657.
KOREN, T. L., POPOV, L. E., DEGTJAREV, K. E., KOVA-
LEVSKY,O.P.&MODZALEVSKAYA, T. L. 2003.
Kazakhstan in the Silurian. In:L
ANDING,E.&
JOHNSON, M. E. (eds) Silurian Lands and Seas:
Palaeogeography Outside of Laurentia. New York
State Museum, Bulletin, 493, 323–344.
KUMP, L. R. 2002. Reducing uncertainty about carbon
dioxide as a climate driver. Nature,419, 188– 190.
KUMP,L.R.&ARTHUR, M. A. 1999. Interpreting
carbon-isotope excursions: carbonates and organic
matter. Chemical Geology,161, 181– 198.
KUMP, L. R., ARTHUR, M. A., PATZKOWSKY, M. E.,
GIBBS, M. T., PINKUS,D.S.&SHEEHAN,P.M.
1999. A weathering hypothesis for glaciation at high
atmospheric pCO
2
during the Late Ordovician.
Palaeogeography, Palaeoclimatology, Palaeoecol-
ogy,152, 173– 187.
LEHERON, D. P., SUTCLIFFE, O. E., W HITTINGTON,
R. J. & CRAIG, J. 2005. The origins of glacially
related soft-sediment deformation structures in Upper
Ordovician glaciogenic rocks: implication for ice
sheet dynamics. Palaeogeography, Palaeoclimatol-
ogy, Palaeoecology,218, 75– 103.
LEGGETT, J. K. 1980. British Lower Paleozoic black
shales and their palaeo-oceanographic significance.
Journal of the Geological Society, London,137,
139–156.
LEGGETT, J. K. 1987. The Southern Uplands as an accre-
tionary prism; the importance of analogues in recon-
structing palaeogeography. Journal of the Geological
Society, London,144, 737– 752.
LEGGETT, J. K., MCKERROW, W. S., C OCKS,L.R.M.&
RICKARDS, R. B. 1981. Periodicity in the early
Palaeozoic marine realm. Journal of the Geological
Society,138, 167– 176.
LEGGETT, J. K., MCKERROW, W. S., MORR IS, J. H.,
OLIVER,G.J.H.&PHILL IPS, W. E. A. 1979. The
north-western margin of the Iapetus Ocean. In:
HARRIS, A. L., HOLLAND,C.H.&LEAKE,B.E.
(eds) The Caledonides of the British Isles; reviewed.
Special Publication of the Geological Society,
London, 8, 499–512.
LEGRAND, P. 1969. Description de Westonia chudeaui sp.
nov. Brachiopode inarticule
´de l’Adrar mauritanien
(Sahara occidental). Bulletin de la Socie
´te
´Ge
´ologique
de France,11, 251–256.
LEGRAND, P. 1970. Les couches a
`Diplograptus du Tassili
de Tarit (Ahmet, Sahara alge
´rien). Bulletin de la
Socie
´te
´d’Histoire Naturelle de Afrique du Nord,60,
3–58.
LEGRAND, P. 1985. Lower Palaeozoic rocks of Algeria.
In:H
OLLAND, C. H. (ed.) Lower Palaeozoic of North-
Western and West-Central Africa. Lower Palaeozoic
Rocks of the World,4, 5–89, John Wiley & Sons,
Chichester.
LOCKLAIR,R.E.&LERMANN, A. 2005. A model of
Phanerozoic cycles of carbon and calcium in the
global ocean: Evaluation and constraints on ocean
chemistry and input fluxes. Chemical Geology,217,
113–126.
LOWENSTAM, H. A. 1961, Mineralogy,
18
O/
13
C ratios,
and strontium and magnesium contents of recent and
A. A. PAGE ET AL.152

fossil brachiopods and their bearing on the history of
oceans. Journal of Geology,69, 241– 260.
LOYDELL, D. K. 1994. Early Telychian changes in grap-
toloid diversity and sea level. Geological Journal,
29, 355– 368.
LOYDELL, D. K. 1998. Early Silurian sea-level changes.
Geological Magazine,135, 447– 471.
LOYDELL,D.K.&NESTOR, V. 2005. Integrated grapto-
lite and chitinozoan biostratigraphy of the upper Tely-
chian (Llandovery, Silurian) of the Ventspils D-3 core,
Latvia. Geological Magazine,142, 369– 376.
LOYDELL, D. K., KALJO,D.&MA
¨NNIK, P. 1998. Inte-
grated biostratigraphy of the lower Silurian of the Ohe-
saare core, Saaremaa, Estonia. Geological Magazine,
135, 769– 783.
LOYDELL, D. K., MA
¨NNIK,P.&NESTOR, V. 2003. Inte-
grated biostratigraphy of the lower Silurian of the
Aizpute-41 core, Latvia. Geological Magazine,140,
205–229.
LUDVIGSON, G. A., WITZKE, B. J., GONZALEZ, L. A.,
CARPENTER, S. J., SCHNEIDER,C.L.&HASIUK,
F. H. 2004. Late Ordovician (Turinian–Chatfieldian)
carbon isotope excursions and their paleoceanographic
significance. Palaeogeography, Palaeoclimatology,
Palaeoecology,210, 187– 214.
LU
¨NING, S., CRAIG, J., LOYDELL, D. K., STORCH ,P.&
FITCHES, B. 2000. Lower Silurian ‘hot shales’ in North
Africa and Arabia: regional distribution and deposi-
tional model. Earth Science Reviews,49, 21– 200.
LU
¨NING, S., LOYDELL, D. K., S
ˇTORCH, P., SHAHIN,Y.
&C
RAIG, J. 2006. Origin, sequence stratigraphy and
depositional environment of an upper Ordovician
(Hirnantian) deglacial black shale, Jordan – Discus-
sion. Palaeogeography, Palaeoclimatology, Palaeoe-
cology,230, 352– 355.
LU
¨NING, S., SHAHIN, Y. M., L OYDELL, D. K., AL-RABI,
H. T., MASRI, A., TARAWNEH,B.&KOLONIC,S.
2005. Anatomy of a world-class source rock: distri-
bution and depositional model of Silurian organic-rich
shales in Jordan and implications for hydrocarbon
potential. AAPG Bulletin,89, 1397– 1427.
MAACK, R. 1957. U
¨ber Vereisungsperioden und Verei-
sungsspuren in Brasilien. Geologische Rundschau,
45, 547– 595.
MABILLARD,J.E.&ALDRIDGE, R. J. 1985. Microfossil
distribution across the base of the Wenlock Series in
the type area. Palaeontology,28, 89– 100.
MA
¨NNIL,R.&MEIDLA, T. 1994. The Ordovician System
of the East European Platform (Estonia, Latvia,
Lithuania, Belorussia, parts of Russia, Ukraine and
Moldova). In:W
EBBY, B. D., ROSS,R.J.&ZHEN,
Y. Y. (eds) The Ordovician System of the East
European Platform and Tuva (southeastern Russia).
IUGS Publication, 28, 1– 52.
MARSHALL, J. D., BRENCHLEY, P. J., MASON, P.,
WOLFF, G. A., ASTINI, R. A., HINTS,L.&
MEIDLA, T. 1997. Global carbon isotopic events
associated with mass extinction and glaciation in the
late Ordovician. Palaeogeography, Palaeoclimatol-
ogy, Palaeoecology,132, 195– 210.
MARTIN, R. E. 1995. Cyclic and secular variation in
microfossil biomineralization – clues to the biogeo-
chemical evolution of Phanerozoic oceans. Global
and Planetary Change,11, 1– 23.
MCCANN,A.M.&KENNEDY, M. J. 1974. A probable
glacio-marine deposit of Late Ordovician–Early Silur-
ian age from the north central Newfoundland Appala-
chian Belt. Geological Magazine,111, 549– 564.
MCCLURE, H. A. 1978. Early Palaeozoic glaciation in
Arabia. Palaeogeography, Palaeoclimatology,
Palaeoecology,25, 315– 326.
MCKERROW, W. S., MAC NIOCIALL,C.&DEWEY,J.F.
2000. The Caledonian Orogeny redefined. Journal of
the Geological Society, London,157, 1149– 1154.
MELCHIN, M. J. 1989. Llandovery graptolite biostratigra-
phy and paleobiogeography, Cape Phillips Formation,
Canadian Arctic Islands. Canadian Journal of Earth
Sciences,26, 1726–1746.
MELCHIN,M.J.&HOLMDEN, C. 2006. Carbon isotope
chemostratigraphy in Arctic Canada: Sea-level
forcing of carbonate platform weathering and impli-
cations for Hirnantian global correlation. Palaeogeo-
graphy, Palaeoclimatology, Palaeoecology,234,
186–200.
MELCHIN, M. J., COOPER,R.A.&SADLER, P. M. 2004.
The Silurian Period. In:G
RADSTEIN, F. M., OGG,J.
G. & SMITH, A. G. (eds) A Geologic Time Scale
2004. Cambridge University Press, Cambridge,
165–187.
MERINO, D. 1991. Primer registro de conodontos siluricos
en Bolivia. Revisita Tecnica de YPFB,12, 271–274.
MEYBECK, M. 1982. Carbon, nitrogen, and phosphorus
transport by world rivers. American Journal of
Science,282, 401– 450.
MILLER,M.A.&MELVIN, J. 2005. Significant new bios-
tratigraphic horizons in the Qusaiba member of the
Silurian Qalibah formation of central Saudi Arabia,
and their sedimentologic expression in a sequence stra-
tigraphic context. GeoArabia,10, 49– 92.
MONTAN
˜EZ, I. P. 2002. Biological skeletal carbonate
records changes in major-ion chemistry of
paleo-oceans. PNAS,99, 15852– 15854.
MU, E.-Z. 1988. The Ordovician– Silurian boundary in
China. Bulletin of the British Museum (Natural
History),43, 117– 131.
MULLINS,G.L.&ALDRIDGE, R. J. 2004. Chitinozoan
biostratigraphy of the basal Wenlock Series (Silurian)
global stratotype section and point. Palaeontology,47,
745–773.
MUNNECKE, A., SAMTLEBEN,C.&BICKERT, T. 2003.
The Ireviken event in the Lower Silurian of Gotland,
Sweden – relation to similar Palaeozoic and Protero-
zoic events. Palaeogeography, Palaeoclimatology,
Palaeoecology,195, 99– 124.
NIELSEN, A. T. 2003a. Ordovician sea level changes: a
Baltoscandinavian perspective. In:W
EBBY, B. D.,
PARIS, F., DROSER,M.&PERCIVAL, I. G. (eds)
The Great Ordovician Biodiversification Event.
Columbia University Press, New York, 84–93.
NIELSEN, A. T. 2003b. Late Ordovician sea level changes:
evidence of Caradoc glaciations? Geophysical
Research Abstracts, European Geosciences Union,5.
NO
˜LVAK, J., HINTS,O.&MA
¨NNIK, P. 2006. Ordovician
timescale in Estonia: recent developments. Proceed-
ings of the Estonian Academy of Sciences, Geology,
55, 95–108.
OLDROYD, D. R. 1990. The Highland Controversy: Con-
structing Geological Knowledge through Fieldwork in
BLACK SHALES IN THE EARLY PALAEOZOIC ICEHOUSE 153

Nineteenth-Century Britain. Science and Its Concep-
tual Foundations series. Chicago University Press,
Chicago.
ORR, P. J. 2001. Bathymetric Indicators. In:BRIGGS,
D. E. G. & CROWTHER, P. R. (eds) Palaeobiology
II. Blackwell Science, Oxford, 479–482.
PANCOST, R. D., FREEMAN, K. H., PATZKOWSKY, M. E.,
WAVREK,D.&COLLISTER, J. W. 1998. Molecular
indicators of redox and marine photoautotroph compo-
sition in the late Middle Ordovician of Iowa, USA.
Organic Geochemistry,29, 1649– 1662.
PARRISH, J. T. 1982. Upwelling and petroleum source
beds, with reference to the Palaeozoic. AAPG Bulletin,
66, 750–774.
PATZKOWSKY, M. E., SLUPIK, L. M., ARTHUR, M. A.,
PANCOST,R.D.&FREEMAN, K. H. 1997. Late
Middle Ordovician environmental change and extinc-
tion: harbinger of the Late Ordovician or continuation
of Cambrian patterns? Geology,25, 911–914.
PHILLIPOT,A.&RO BARDET, M. 1971. Nouvelles
donnees sur les formations siluriennes de Domfront
(Orne). Bulletin de la Socie
´te
´Ge
´ologique et Mineralo-
gique de Bretagne, Serie C,3, 41–47.
PICKERILL, R. K., PAJARI,G.E.&CURRIE, K. L. 1979.
Evidence of Caradocian glaciation in the Davidsville
Group of north-eastern Newfoundland. Current
Research, Part C, Papers of the Canadian Geological
Survey,79, 67– 72.
PORE˛BSKA,E.&SAWłOWICZ, Z. 1997. Palaeo-
ceanographic linkage of geochemical and graptolite
events across the Silurian–Devonian boundary in
Bardzkie Mountains (Southwest Poland). Palaeo-
geography, Palaeoclimatology, Palaeoecology,132,
343–354.
PRATT, W. T., WOODHALL,D.G.&HOWELLS,M.F.
1995. Geology of the country around Cadair Idris.
Memoir for 1:50,000 Geological Sheet 149 (England
and Wales). London, HMSO.
RAHMSTORF, S. 2002. Ocean circulation and
climate during the past 120,000 years, Nature,419,
207–214.
RAYMO, M. E. 1991. Geochemical evidence supporting
T.C. Chamberlin’s theory of glaciation. Geology,19,
344–347.
RICKARDS, R. B. 1970. The Llandovery (Silurian) grapto-
lites of the Howgill Fells, Northern England. Palaeon-
tographical Society Monographs,123, 1– 108.
RICKARDS,R.B.&WOODCOCK, N. H. 2005. Stratigra-
phical revision of the Windermere Supergroup (Late
Ordovician– Silurian) in the southern Howgill Fells,
NW England. Proceedings of the Yorkshire Geologi-
cal Society,55, 263– 285.
RIDGWELL, A. 2005. A Mid Mesozoic Revolution in the
regulation of ocean chemistry. Marine Geology,217,
339–357.
ROBARDET, M., HENRY, J. L., NION, J., PARIS,F.&
PILLET, J. 1972. La Formation du Pont-de-Caer
(Caradocien) dans les synclinaux de Domfront et de
Sees (Normandie). Socie
´te
´Ge
´ologique du Nord,
Annales,92, 117– 137.
ROCHA-CAMPOS, A. C. 1981. Early Palaeozoic Iapo For-
mation of Parana
´, Brazil. In:H
AMBREY,M.J.&
HARLAND, W. B. (eds) Earth’s pre-Pleistocene
Glacial Record. Cambridge University Press, Cam-
bridge, 908–909.
ROHLING,E.J.&GIESKES, W. W. C. 1989. Late Qua-
ternary changes in Mediterranean Intermediate Water
density and formation rate. Paleoceanography,4,
531–545.
ROHLING, E. J. 1994. Review and new aspects concerning
the formation of Mediterranean sapropels, Marine
Geology,122, 1– 28.
RONG, J.-Y., BOUCOT, A. J., S U, Y.-Z. & STRUSZ,D.L.
1995. Biogeographical analysis of Late Silurian bra-
chiopod faunas, chiefly from Asia and Australia.
Senckenberg Lethaea,28, 39– 60.
RONOV, A. B., KHAIN, V. E., BALUKHOVSKY,A.N.&
SESLAVINSKY, K. B. 1980. Quantitative analysis of
Phanerozoic sedimentation. Sedimentary Geology,
25, 311– 325.
ROSS,C.A.&ROSS, J. R. P. 1995. North American Ordo-
vician depositional sequences and correlations. In:
COOPER,J.D.&FINNEY, S. C. (eds) Ordovician
Odyssey. SEPM, Pacific Section, Fullerton California,
309–313.
ROSS,C.A.&ROSS, R. P. 1996. Silurian sea-level
fluctuations. In:W
ITZKE, B. J., LUDVIGSON,G.A.
&D
AY, J. (eds) Paleozoic Sequence Strati-
graphy: Views from the North American Craton.
Geological Society of America, Special Paper, 306,
187–192.
ROTHMAN, D. H., HAYES,J.M.&SUMMONS,R.E.
(2003) Dynamics of the Neoproterozoic carbon
cycle. Proceedings of the National Academy of
Science,100, 8124– 8129.
ROYER, D. L. 2006. CO
2
-forced climate thresholds during
the Phanerozoic. Geochimica et Cosmochimica Acta,
70, 566– 567.
ROYER, D. L., BERNER, R. A., M ONTAN
˜EZ, I. P.,
TABOR,N.J.&BEERLING, D. J. 2004. CO
2
as a
primary driver of Phanerozoic climate. GSA Today,
14, 4– 10.
RUDWICK, M. J. S. 1985. The great Devonian contro-
versy: The shaping of scientific knowledge among
gentlemanly specialists. Science and its conceptual
foundations series. Chicago University Press,
Chicago.
RUST, I. C. 1981. Early Palaeozoic Pakhuis Tillite, South
Africa. In:H
AMBREY,M.J.&HARLAND, W. B. (eds)
Earth’s pre-Pleistocene Glacial Record. Cambridge
University Press, Cambridge, 113– 117.
SALTZMAN, M. R. 2005. Phosphorus, nitrogen, and the
redox evolution of the Paleozoic oceans. Geology,
33, 573– 576.
SALTZMAN,M.R.&YOUNG, S. A. 2005. A long-lived
glaciation in the Late Ordovician? Isotopic and
sequence stratigraphic evidence from western Lauren-
tia. Geology,33, 109– 112.
SAMTLEBEN, C., MUNNECKE, A., BICKERT,T.&
PA
¨TZOLD, J. 2001. Shell succession, assemblage and
species dependent effects on the C/O-isotopic compo-
sition of brachiopods: examples from the Silurian of
Gotland. Chemical Geology,175, 61– 107.
SAMTLEBEN, C., MUNNECKE, A., BICKERT,T.&
PA
¨TZOLD, J. 1996. The Silurian of Gotland
(Sweden): Facies interpretation based on stable
A. A. PAGE ET AL.154

isotopes in brachiopod shells. Geologische Rundschau,
85, 278– 292.
SCHENCK, P. E. 1972. Possible Late Ordovician glacia-
tion of Nova Scotia. Canadian Journal of Earth
Science,9, 95– 107.
SCHENCK,P.E.&LANE, T. E. 1981. Early Paleozoic
tillite of Nova Scotia, Canada. In:H
AMBREY,M.J.
&H
ARLAND, W. B. (eds) Earth’s pre-Pleistocene
Glacial Record. Cambridge University Press,
Cambridge, 707– 710.
SCHOFIELD, D. I., DAVIES,J.R,WATERS, R. A.,
WILBY, P. R., WILLIAMS,M.&WILSON, D. 2004.
Geology of the Builth Wells district – a brief expla-
nation of the geological map. Sheet explanation of
the British Geological Survey. 1:50000 sheet 196
Builth Wells (England & Wales). Keyworth,
Nottingham, British Geological Survey.
SCOTESE,C.R.&MCKERROW, W. S. 1991. Ordovician
plate tectonic reconstructions. Canadian Geological
Survey Paper,90– 99, 271 – 282.
SCRIVNER, A. E., VANCE,D.&ROHLING, E. J. 2004.
New neodymium isotope data quantify Nile involve-
ment in Mediterranean anoxic episodes. Geology,32,
565–568.
SECORD, J. A. 1986. Controversy in Victorian geology:
the Cambrian– Silurian dispute. Princeton University
Press, Princeton.
SHACKLETON, N. J. 2000. The 100,000-year ice-age cycle
identified and found to lag temperature, carbon
dioxide, and orbital eccentricity. Science,289,
1897– 1902.
SHAVIV, N. J. 2002. Cosmic ray diffusion from the galactic
spiral arms, iron meteorites, and a possible climatic
connection. PhysicalReview Letters,89,art. no. 051102.
SHAVIV,N.J.&VEIZER, J. 2003. Celestial driver of Pha-
nerozoic climate?, GSA Today,13, 4– 10.
SHIELDS, G. A., CARDEN, G. A. F., VEIZER, J., MEIDLA,
T., RONG, J.-Y. & LI, R.-Y. 2003. Sr, C and O isotope
geochemistry of Ordovician brachiopods: a major iso-
topic event around the Middle– Late Ordovician tran-
sition. Geochimica et Cosmochimica Acta,67,
2005– 2025.
SHUSTER, D. L., EHLERS, T. A., RUSMOREN,M.E.&
FARLEY, K. A. 2005. Rapid glacial erosion at
1.8 Ma revealed by
4
He/
3
He thermochronometry.
Science,310, 1668– 1670.
SIEGENTHALER, U., STOCKER, T., MONNIN,E.ET AL.
2005. Stable carbon cycle-climate relationship during
the late Pleistocene. Science,310, 1313– 1317.
SKEVINGTON, D. 1974. Controls influencing the compo-
sition and distribution of Ordovician graptolite pro-
vinces. In:R
ICKARDS, R. B., JACKSON,D.E.&
HUGHES, C. P. (eds) Graptolite Studies in Honour of
O. M. B. Bulman. Special Papers in Palaeontology,
13, 59– 73.
SMITH,A.G.&PICKERING, K. T. 2003. Oceanic gate-
ways as a critical factor to initiate icehouse Earth.
Journal of the Geological Society,160, 337–340.
S
ˇTORCH, P. 1986. Ordovician –Silurian boundary in the
Prague Basin (Barrandian area, Bohemia). Sbornı
´k
Geologicky
´ch ve
ˇd, Geologie,41, 69–103.
S
ˇTORCH, P. 1994. Graptolite biostratigraphy of the Lower
Silurian (Llandovery and Wenlock) of Bohemia.
Geological Journal,29, 137– 165.
S
ˇTORCH,P.&MERGL, M. 1989. Kra
´lodvor/Kosov
boundary and the late Ordovician environmental
changes in the Prague Basin (Barrandian area,
Bohemia). Sbornı
´k Geologicky
´ch Ve
ˇd, Geologie,44,
117–153.
STRACHAN, R. A. 2000. Mid-Ordovician to Silurian sedi-
mentation and tectonics on the northern active margin
of Iapetus. In:W
OODCOCK,N.H.&STRACHAN,R.
A. (eds) Geological History of Britain Ireland. Black-
well, Oxford, 107–123.
SUA
´REZ-SORUCO, R. 1995. Comentarios sobre la edad de
la Formacion Cancan
˜iri. Revista Tecnica de Yacamien-
tos Petroliferos Fisicales Bolivianos,16, 51–54.
SUTCLIFFE, O. E., DOWDESWELL, J. A., W HITTINGTON,
R. J., THERON,J.N.&CRAIG, J. 2000. Calibrating
the Late Ordovician glaciation and mass extinction
by the eccentricity of Earth’s orbit. Geology,28,
967–970.
SUTCLIFFE, O. E., HARPER, D. A. T., SALEM, A. A.,
WHITTINGTON,R.J.&CRAIG, J. 2001. The develop-
ment of an atypical Hirnantia brachiopod fauna and
the onset of glaciation in the late Ordovician of Gond-
wana. Transactions of the Royal Society of Edinburgh,
Earth Sciences,92, 1–14.
THONG-DZUY, T., BOUCOT, A. J., RONG, J.-Y. & FANF,
Z.-J. 2001. Late Silurian marine shelly fauna
of Central and Northern Vietnam. Geobios,34,
315–338.
TOBIN, K. J., BERGSTRO
¨M,S.M.&DELAGARZA,P.
2005. A mid-Caradocian (453 Ma) drawdown in
atmospheric pCO
2
atmospheric CO
2
without ice
sheet development? Palaeogeography, Palaeoclima-
tology, Palaeoecology,226, 187– 204.
TOGHILL, P. 1968. The graptolite assemblages and zones
of the Birkhill Shales (Lower Silurian) at Dobb’s Linn.
Palaeontology,11, 654– 678.
TORSVIK, T. H. 1998. Palaeozoic palaeogeography: A
North Atlantic viewpoint. GFF,120, 109– 118.
TUCKER,M.E.&REID, P. C. 1973. The sedimentology
and context of late Ordovician glacial marine sedi-
ments from Sierra Leone, West Africa. Palaeo-
geography, Palaeoclimatology, Palaeoecology,13,
289–307.
UNDERWOOD, C. J. 1992. Graptolite Preservation and
Deformation. Palaios,7, 178– 186.
UNDERWOOD, C. J., CROWLEY, S. F., M ARSHALL,J.D.
&B
RENCHLEY, P. J. 1997. High resolution carbon
isotope stratigraphy of the basal Silurian stratotype
(Dob’s Linn, Scotland) and its global correlation.
Journal of the Geological Society, London,154,
709–718.
UNDERWOOD, C. J., DEYNOUX,M.&GHIENNE, J.-F.
1998. High Palaeolatitude (Hodh, Mauritania)
recovery of graptolite fauna after the Hirnantian
(end Ordovician) extinction event. Palaeogeography,
Palaeoclimatology, Palaeoecology,142, 91– 105.
VEIZER, J., BRUCKSCHEN,P.&PAWELLEK,F.ET AL.
1997. Oxygen isotope evolution of Phanerozoic
seawater. Palaeogeography, Palaeoclimatology,
Palaeoecology,132, 159– 172.
VEIZER, J., GODDERIS,Y.&FRANC¸OIS, L. M. 2000.
Evidence for decoupling of atmospheric CO
2
and
global climate during the Phanerozoic eon. Nature,
408, 698– 701.
BLACK SHALES IN THE EARLY PALAEOZOIC ICEHOUSE 155

VELBEL, M. A. 1993. Temperature dependence of silicate
weathering in nature: How strong a negative feedback
on long-term accumulation of atmospheric CO
2
and
global greenhouse warming? Geology,21, 1059–1062.
VERNIERS,J.&VANDENBROUCKE, T. R. A. 2006.
Chitinozoan biostratigraphy in the Dob’s Linn Ordovi-
cian–Silurian GSSP, Southern Uplands, Scotland.
GFF,128, 195– 202.
WALKER, L. J., WILKINSON,B.H.&IVANY, L. C. 2002.
Continental drift and Phanerozoic carbonate accumu-
lation in shallow-shelf and deep-marine settings.
Journal of Geology,110, 75– 87.
WHITE, D. E., BARRON, H. F., BARNES,R.P.&
LINTERN, B. C. 1991. Biostratigraphy of late Llan-
dovery (Telychian) and Wenlock turbiditic sequences
in the SW Southern Uplands, Scotland. Transactions
of the Royal Society of Edinburgh, Earth Sciences,
82, 297–322.
WIGNALL, P. B. 1991. Model for transgressive black
shales? Geology,19, 167– 170.
WIGNALL, P. B. 1994. Black Shales. Clarendon Press,
Oxford.
WIGNALL,P.B.&MAYN ARD, J. R. 1993. The sequence
stratigraphy of transgressive black shales. In:K
ATZ,
B. J. & PRATT, L. (eds) Source Rocks within a
Sequence Stratigraphic Framework. American
Association of Petroleum Geologists Studies in
Geology, 37, 35– 47.
WILLIAMS, M., DAVIES, J. R., W ATERS, R. A.,
RUSHTON,A.W.A.&WILBY, P. R. 2003. Stratigra-
phical and palaeoecological importance of Caradoc
(Upper Ordovician) graptolites from the Cardigan
area, southwest Wales. Geological Magazine,140,
549–571.
WILLIAMS, M., RUSHTON, A. W. A., WOOD, B., FLOYD,
J. D., SMITH,R.&WHEATLEY, C. 2004. A revised
graptolite biostratigraphy for the lower Caradoc
(Upper Ordovician) of southern Scotland. Scottish
Journal of Geology,40, 97– 114.
WILLIAMS, S. H. 1982. The Late Ordovician fauna of the
anceps bands at Dob’s Linn, southern Scotland. Geo-
logica et Palaeontologica,16, 29– 56.
WOODCOCK, N. H. 1990. Sequence stratigraphy of the
Palaeozoic Welsh Basin. Journal of the Geological
Society, London,147, 537– 547.
WOODCOCK, N. H. 2000. Late Ordovician to Silurian
evolution of Eastern Avalonia during convergence
with Laurentia. In:W
OODCOCK,N.H.&STRACHAN,
R. A. (eds) Geological History of Britain and Ireland.
Blackwell, Oxford.
WOODCOCK, N. H., BUTLER, A. J., D AVIES,J.R.&
WATERS, R. A. 1996. Sequence stratigraphical
analysis of late Ordovician and early Silurian deposi-
tional systems in the Welsh Basin: a critical assess-
ment. In:H
ESSELBO,S.P.&PARKINSON,D.N.
(eds) Sequence Stratigraphy in British Geology.
Special Publication of the Geological Society of
London, 103, 197– 208.
YAPP,C.J.&POTHS, H. 1992. Ancient atmospheric CO
2
pressures inferred from natural goethites. Nature,355,
342–344.
YOLKIN, E. A., SENNIKOV, N. V., BAKHAREV, N. K.,
IXOKH,N.G.&YAZIKOV,A.YU. 1997. Periodicity
of deposition in the Silurian and relationships of
global geological events in the middle Paleozoic of
the southwestern margin of the Siberian continent.
Geologiya i Geofizika,38, 596– 607.
YOUNG, G. M. 1981. Early Palaeozoic tillites of the north-
ern Arabian Peninsula. In:H
AMBREY,M.J.&
HARLAND, W. B. (eds) Earth’s Pre-Pleistocene
Glacial Record. Cambridge University Press,
Cambridge, 338– 340.
YOUNG, S. A., SALTZMAN,M.R.&BERGSTROM,S.A.
2005. Upper Ordovician (Mohawkian) carbon isotope
stratigraphy in Eastern and Central North America:
regional expression of a perturbation of the global
carbon cycle. Palaeogeography, Palaeoceanography,
Palaeoclimatology,222, 53– 76.
ZACHOS, J., PAGANI, M., S LOAN, L., THOMAS,E.&
BILLUPS, K. 2001. Trends, rhythms, and aberrations
in global climate 65 Ma to present. Science,292,
686–693.
ZALASIEWICZ, J. A. 2001. Graptolites as constraints on
models of sedimentation across Iapetus: a review. Pro-
ceedings of the Geologists Association,112, 237–251.
ZALASIEWICZ,J.A.&WILLIAMS, M. 1999. Graptolite
biozonation of the Wenlock Series (Silurian) of the
Builth Wells district, central Wales. Geological Maga-
zine,136, 263– 283.
ZALASIEWICZ, J. A., RUSHTON,A.W.A.&OWEN,A.
W. 1995. Late Caradoc graptolitic faunal gradients
across the Iapetus Ocean. Geological Magazine,132,
611–617.
ZALASIEWICZ, J., WILLIAMS, M., M ILLER, M., PAGE,
A. & BLACKETT, E. 2007. Early Silurian (Lland-
overy) graptolites from central Saudi Arabia: first
documented record of Telychian faunas from the
Arabian Peninsula. GeoArabia,12, in press.
ZHANG, S., BARNES,C.R.&JOWETT, D. M. S. 2006.
The paradox of the global standard Late Ordovi-
cian–Early Silurian sea level curve: Evidence from
conodont community analysis from both Canadian
Arctic and Appalachian margins. Palaeogeography,
Palaeoclimatology, Palaeoecology,236, 246– 271.
A. A. PAGE ET AL.156