6%0
LETIERSTO
NATURE
NATURE
VOL. 307 16
FEBRUARY
1984
Table 2
REE
by
is
otope
dilution
Sample
La
Ce
Nd
Sm
42162
CPX·
0.
90
3.60 5.33
2.00
42158
CPX
0.59
2.66 3.
57
1.36
42157
CPX
0.83
3.73 5.34 2.15
42156CPX
2.
94
3.99
1.54
42154CPX
0.44
2.
06
2.77
1.08
42149CPX
0.79
2.77
3.790
1.48
42144CPX
0.79
3.63
4.79
1.83
42142CPX
3.78 4.
57
1.75
41596CPX
0.77
3.26
4.
360
1.65
From ref. 15.
Error
s 2 s.e.
are
1 %.
CPX,
clinopyroxene.
Table 2.
The
REE
patterns
are
parallel, with variable, small,
negative
Eu
anomalies. Because the patterns
are
parallel the
total
REE
content
(t
REE)
of
the
pyroxenes can be expressed
in terms of
one
element, Nd, for example.
The
variation of Nd
through unit 9 (Fig. 2)
is
remarkably similar
to
87Sr/86Sr.
Although Nd can increase in magma during fractionation,
8
7Sr/R
6
Sr
cannot, consequently, these variations represent
the
addition of a high Nd, high 87Sr/86Sr contaminant. This addition
is
seen
in
the gabbros
at
the
top
of unit 9 and coincides with
Received 25 May; accepted 18 November 1983.
I. DePaol
o,
D. & Wasserburg, G.
Oe
ochim. cosmochim. A cla 43,
61~27
(1979).
2. James, D.
E.
Banh
planet. Sci. Lell.
57,47-62
(19
82).
3.
DePaolo, D.
Eanh
planel. Sci. UII. 53,
189-20
2 (1982
).
4. M
oo
rbath, S. & Welke, H.
Eanh
planet. Sci. LetL 5,
217-230
(1969
).
5.
M
oo
rbath,
S.
& Thompson, R. N. J. Pelfol. 21, 295
-321
(1980).
6.
Dickin, A. P. J. Pelfol.
2%,155-189
(1980).
7. Brown, G. M. Phil. Trans.
R.
Soc. B%40,
I-5
3 (1956
).
Oxygen isotope calibration of the onset
of ice-rafting and history of
glaciation in the North Atlantic region
N. J. Shackleton*, J. Backman,
H.
Zimmerman,
D. V. Kent,
M.
A. Hall, D. G. Roberts, D. Schnitker,
J. G. Baldauf, A. Desprairies, R. Homrighausen,
P. Huddleston,
J. B. Keene, A. J. Kaltenback,
K.
A. O. Krumsiek, A. C. Morton,
J. W. Murray & J. Westberg-Smith
We report here that
DSDP
Site SS2A, cored with the hydraulic
piston corer on the west flank of
RockaU Bank, recovered an
undisturbed sequence of alternating white deep-sea carbonate
oozes and dark-coloured layers that are rich in glacial debris.
Oxygen isotope analysis of the sequence together with detailed
nannofossil and palaeomagnetic stratigraphy shows that the first
major horizon of ice-rafting occurred
at
about 2.4 Myr, and
was preceded
by
a minor pulse of ice-rafting
at
about 2.5 Myr.
The carbon isotope record shows that the site has been bathed
by
a water mass of similar characteristics to present-day North
Atlantic deep water
at
least since 3.5 Myr.
Despite the concentration of members of
the
deep-sea drilling
community around the
North
Atlantic, surprisingly few
Deep
Sea Drilling Project sites have
been
drilled
that
are
suitable for
even a cursory study of
North
Atlantic palaeoenvironments.
Until recently the best sites in which
to
study
the
onset of
glaciation around the
North
AtlantiC were from
DSDP
Leg 12,
sites 111 and 116,
in
which Berggren! studied the stratigraphic
• Authors' addresse
s:
Godwin Laboratory for Quaternary Research, Free School
lane,
Cambridge C
B2
3RS
(N.l
.S.); Department
of
Geology,
Un
iversity
of
Stockholm, S.
1069O
Stockholm, Sweden (J.B.); Department
of
Civil Engineering, Union College, Schenectady, New
York
12308
, USA (H.Z.); Lamond Doherty Geological Observatory
of
Columbia University,
Palisades, New York 10964, USA (N.J.s., D.V.K.); British Petroleum (Overseas) Development
Ltd, Britannic House, Moorgate,
London
ECI,
UK
(D.G.R.); Department
of
Oceanography,
University
of
Maine, Walpole, Maine 04573, USA (D.S.); United States Geological Survey, 345
Middlefield Road, Menlo Park, California 94025, USA (J.G.B.): Laboratoire
de
Geochimie des
0028-{)836/ 84/
070620-04S01
.
00
Eu
Od
Dy
Er
Yb
0.69
2.
60
2.75 1.37 1.00
0.47 1.79
1.78
0.
90
0.6
8
0.75 2.92
2.99 1.54 1.20
0.48
0.81
0.38
1.42 1.45 0.71 0.55
0.49
1.95
0.94
0.83
0.58
2.42 1.29
1.000
0.56
2.
29
1.29
0.97
0.52 2.
03
1.19
0.94
fractionation.
The
peridotites of unit 9 also seem
to
be contami-
nated
. However, this contamination could be the result of
magma mixing between the fresh magma influx
and
any contami-
nated
magma left in
the
chamber
.
I
thank
Drs
M.
F. Thirlwall, R.
A.
Cliff
and
B.
M.
Wilson for
helpful discussion
and
criticism and
Dr
E. Condliffe for advice
on
the
microprobe. Isotope work at Leeds
is
funded by the
Royal
Society and
NERC,
and
this work was performed during
the
tenure
of a
NERC
studentship.
8.
Huppert
, H. E. & Sparks, R. S.
J.
Co
nlf
.
Mi
ner.
Pe
lf
ol. 75, 27
9-2
89 (1980
).
9.
Roede
r, P. L. & Emslie. R. F. Conlr. Miner. Petrol. 29, 275-
289
(1
970).
10.
Dunham
, A. C. & Wadsworth, W. 1. Mi
n.
Mag.
42
, 347
-3
56
(1979).
11. Vollmer, R.,
et
al
. Oeolherm.
Res
. 11, 317
-327
(1
98
1).
1
2.
Sneeringer, M. &
Hart
, S.
EOS
59, 402
(1
97
8).
13. Ta
yl
or. H. P. & Forester. R. W.
1.
Pelfol.
20
, 3
55-419
(1979).
14. St
o,
ch, H. G., Carlson, R. W. & Lugmair, G. W. Ea
nh
. planet. Sci. Lett. 47, 26
3-
271
(1980
).
1
5.
Thirlwall. M. F., Chern. Oeol.
35,
155
-1
66
(1
982
).
Fig. 1 Location
map
for Site
552A
(56°
02
.
56'
N, 23°13.38' W,
2,311 m water
depth
.
position of the earliest ice-rafted debris in the region
and
esti-
mated
its age
at
3.0
Myr
(Backman
2
re-examined this material
and obtained a younger estimate of 2.5 Myr). Sediments from
the Leg 12 sites were extensively disturbed by rotary drilling,
and
not
suitable for detailed analysis of
the
palaeoenvironmental
record. During Leg 81,
the
Glomar Challenger used
the
newly
developed hydraulic piston
corer
(HPC)3 with considerable suc-
cess, and in particular recovered a section from
the
west flank
of
Rockall
Bank
(Fig.
1)
that
is
largely undisturbed .
Rocbel
Sidimentaires, University
of
Paris
XI
, 91405 Orsay. France
(A.D.);
Laboratorium filr
Erdolgewinnung, Deutsche Texaco AG,
3101
Wietze,
FRG
(R.H.) ; Georgia Geological Survey,
19
Martin Luther King
Jr
Drive, Atlanta,
Ge
orgia 30334, USA (P.H.); Department
of
Geology,
University
of
Melbourne. Parkville. Victoria 3052, Australia (J.B.); Denver Research Center,
Marathon
Oil Co., 7400 South Broadway, Littleton, Colorado SOl60, USA(A.l.K.); Geologisch
..
Institut, Nussallee
S,
D-53 Bonn I,
FRG
(K
.A.
O.K
.); Institute
of
Geological Sciences, Ring Road,
Halton, Leeds LSI5 STQ,
UK
(A.C.M.);
Dep
artment
of
Geology, University
of
Exeter, Exeter,
UK (J.W.M.): Scripps Institution
of
Oceanograph
y,
La
Jolla, California 92092, USA
(J
.W.·S
.).
© 1984 Macmillan
Journ
als Ltd
© Nature Publishing Group1984
~NA~T~U~R~E~V~0~L~.~30~7~1~6~F~E~BR~U~A~R~Y~19~84~
________
--------LETTERSTONATURE-------------------------------------6~11
1 2 3
4,
5 6 7 8
9"
I ,10,
11
12
CORE
I '
'I
' , 'I ' , 'I ' , 'I I ' , 'I ' ,
'140'
'I ' , 'I ' , 'I
~~--~----~1P------~----~------~----~3p~----~----~,------~-----~~0----~
____
~6pm
(]
81M
eM
J\lJlJ
nf\
o R
MIG
K
GIG
~
I i I I i
.73
.91
1.66
2.47
2.92
3.40
Fig. 2 Magnetic record for Site 552A. Demagnetized inclinations are shown only for apparently undisturbed parts of the cores (the data
from core
11
suggest that some unrecognized disturbance may be present).
The upper 40 m of the section comprises alternating units of
white and dark grey colour. In the white intervals,
the>
150-llin
fraction consists almost exclusively of foraminifera;
in
the dark
intervals angular sand-sized rock fragments predominate
although sufficient benthic foraminifera for isotopic analysis
are present at almost every horizon examined. The dark-
coloured intervals containing ice-rafted material are visually
clear and also show unmistakably
in
a record of carbonate
content (typically between 80 and 90%
in
the white layers and
10-40% in the dark layers).
We have made isotope analyses at 10-cm intervals over the
interval from well below the first ice-rafting layer, through to
the recent sediment. Only core 6 was excluded from analysis
because it was very severely disturbed
by
coring. Three species
were used:
Globocassidulina subglobosa, Uvigerina peregrina
and Planulina wuellerstorfi. In Figs 3 and 4
data
are presented
using the previously established 'correction factors' to relate the
oxygen isotope values obtained for
G.
subglobosa and
P.
wueller-
storfi
to
the equilibrium value which
is
thought
to
be given
by
Uvigerina
Spp.4.5.
Carbon isotope values are related
to
P.
wuel-
lerstorfi
rather than to Uvigerina because this procedure pro-
vides a good estimate of the
l3C
content of ocean dissolved
CO
2
(refs 6, 7).
Samples for palaeomagnetic analysis were taken approxi-
mately every
25
cm, avoiding sediment that showed any sign of
disturbance. Only orientation with respect
to
the vertical axis
was preserved so that the inclination component
is
used
to
infer
polarity at this high-latitude site. The natural remanent mag-
netization (NRM) was measured on a two-axis cryogenic
magnetometerS both before and after alternating-field demag-
netization at lO-30 mT. NRM intensities
are
typically of the
order lO-2 A
m-
I
in the upper 45 m of the section, decreasing
to about
10-
3
A
m-
I
in the lower part. Measured inclination data
are shown
in
Fig.
2.
Inclinations are generally well-grouped near
to the expected dipole value (71
0
,
positive for normal and
negative for reversed) for the latitude of the site, indicating that
a reliable record of the geomagnetic field has been obtained to
the base of the Gauss normal magnetochron. The lower boun-
dary of the Olduvai subchron
is
unfortunately obscured by
disturbance at the break between successive cores, while the
Mammoth subchron
is
unclear, probably lying in a slightly
disturbed section of core 11.
Analyses of the carbonate content of the sediment have also
been made at lO-cm intervals through the upper 59 m of the
section. In this part of the North Atlantic, carbonate content
is
controlled chiefly
by
variations
in
the influx of non-carbonate
material transported
by
ice
9
.
Changing carbonate production
also has a major role, while carbonate dissolution has little
impact at such shallow water depths in the North Atlantic.
The samples taken for stable isotope analysis were also used
for quantitative nannofossil studies; here we show (Fig. 4) only
those nannofossil data which have a bearing
on
the biostrati-
graphical position of the early ice-rafting horizon.
Figure 3 shows the oxygen isotope and carbonate data for
the upper
five
cores, together with the oxygen isotope record
of
core V28-238 (ref.
4)
for comparison. Both records are
plotted linearly with respect
to
depth in sediment between
magnetic reversals. Since it has been established that the sedi-
ment in V28-238 accumulated at a fairly regular rate
lO
, the axis
0.0
0.73
0.118
(I!~~f
~~
~
4.0
C
4.5
~
Q
5.0
w
1001L------------------~----r_------~----~T
!E
40
~
20
r
-3.0
-2.5-
0.0-
0.5-
1~
I I I I I I I I I I I
0.0
0.1
0.2
0.3
0.4 0.5
0.'
0.7
0.' 0.'
1.0
1.1
1.2
AGE. MILLION YEARS
Fig. 3 Top: Oxygen isotope record of Site 552A cores 1-5. The
plotting scale
is
linear between magnetic reversal horizons. Middle:
carbonate content
in
the same interval; low carbonate content
implies dilution
by
ice-rafted material. Bottom: oxygen isotope
record of Pacific core V28-238 (ref.
4)
for comparison. Vertical
lines indicate the horizons used for time control: top (zero age),
base Bruhnes normal chron (0.73 Myr), base Jaramillo subchron
(0.98 Myr).
is
shown as an approximate age scale calibrated on the basis of
the single magnetic reversal in this core. Thus the comparison
shows that there have been quite marked accumulation rate
variations
in
DSDP 552A. This
is
hardly surprising
in
view of
the lithological variations that are evident. Nonetheless the
oxygen isotope record of Site 552A
is
complete to the extent
of preserving every isotope stage.
In Figure 4, the oxygen isotope data for the lower part of the
sequence are shown and compared with the oxygen isotope data
for piston core V28-1 79. (Figure 4 shows additional measure-
ments that
we
have made
in
core V28-179. Note that the position
© Nature Publishing Group1984
/ill
~--------------------
_________________
LETTERSTONATURE
______________
~N~A~TU~R~E~V~O~L~.~30~7~1~6~FE~B~R~U~A~RY~19~84
2.0
2.8 -
3.0
=
I
U
m
~
4.0
)(
0
i
4.11
a.o
...
a.o
I
100
80
eo
I
40
20
0
2.0
2.8
3.0
«I
...
I
!
u
4.0
C
i
4.11
a.o
...
1
...
1A7
D.P.
D.B. D.S.
D.P.
D.T.
I
Fig. 4
Top
: Oxygen isotope record
of Site
552A
cores
7-12
.
The
plotting
scale is linear between magnetic
reversal horizons. Middle: carbonate
content
in same interval. Bottom:
oxygen isotope record in Pacific
core
V28-179
(ref. 11) for comparison.
Vertical lines show horizons used
for time control:
top
Olduvai
normal subchron
(1.66 Myr), base
Matuyama
reversed
chron
(2.47
Myr),
top
Kaena
reversed subchron
(2.92 Myr), base Gauss normal
chron
(3.
40
Myr). Nannofossil
extinction horizons
determined
in
both cores indicated by:
DT,
Dis-
coastertamalis; DS, D. surculus;
DP
,
D. pentaradiatus; DB, D. brouweri .
a.o-r'-~-r~,,-r~,,-.~,,-.~,,-.~,,-.~,,-.~~-'~~-'~~~~~~
1.4 1.8 1.1 1.7 1.1 1.8 2.1
U
IA
2.8 U
i7
U
i.
a:o
a:l
U U
:1.4
U a:1
AGE, MILLION YEARS
of the Gauss-Matuyama boundary was incorrectly printed in
ref. 11; it was
in
fact located at 1,338 cm, not 1,358 cm.) Again,
to facilita.te comparison, the plotting scale for both records has
been adjusted at magnetostratigraphic boundaries as indicated,
so
as to approximate an age scale using a palaeomagnetic time
scale
12
• Obviously the application of this age scale becomes less
accurate far from the magnetic reversals. However, comparison
between the two records
is
good considering the low accumula-
tion rate, and relatively coarse sampling interval in V28-179,
and many small-scale features can be correlated between the
two. The amplitude of variation
is
substantially greater in Site
552A than
in
V28-179. This
is
probably due to the effect of
bioturbation
in
the Pacific core; since the accumulation rate of
core V28-179
was
only about 0.55
cm
kyr-l,
one would expect
climatic extremes lasting only a few thousand years to be severely
degraded by bioturbation.
The series of nannofossil extinction data shown in Fig. 4 have
also been determined quantitatively in core V28-179 (ref. 13)
and their ages estimated. Using these estimates, the average
accumulation rate over the pre-ice-rafting interval in Site 552A
may be estimated at approximately 1.7 cm
kyr-l,
surprisingly
close to the average for the upper part, 1.8 cm kyr-
1
(although
the rate has certainly varied over short intervals during the
intense climatic fluctuations of the past few million years).
Palaeomagnetic stratigraphy and nannofossil stratigraphy in this
site give identical estimates for the age of the first major northern
hemisphere glacial event about 2.37 Myr.
It
is
clear from an examination of Figs 3 and 4 that there
is
a remarkable correspondence between the records of oxygen
isotopic composition in the foraminifera, and carbonate content,
back to about 2.4 Myr. This
is
not particularly surprising since
both are clearly causally related to glaciation. However, it
is
worth noting that
if
the late Pliocene isotopic fluctuations were
attributed to glaciation in Antarctica, one would not expect the
influx of ice-rafted debris during the glaciation at 2.4 Myr to
be
as dramatic as that observed in the last glacial, whereas this
is
what
we
observe in Site 552A. A brief ice-rafting episode,
coinciding with positive
18
0 values,
is
observed at about 2.5 Myr.
Prior to 2.5 Myr, the influx of ice-rafted debris
is
minute and
has little impact on the record of carbonate content. Presumably
such glaciation as may have occurred before that time did not
give rise to extensive calving of icebergs into the North Atlantic.
If
we
compare the isotopic composition of benthonic
foraminifera at the level of the first glacial maximum
at
2.37 Myr
with values
in
glacial levels
in
the upper part of the sequence,
it becomes apparent that the early event represents a truly
glacial interval, with an
ice
volume similar to maxima during
the middle Pleistocene. This must, for instance, be correlative
with a fuily glacial environment in Britain although obviously
the actual ice-margin positions cannot be determined from these
data.
Other
data relating to a major climatic threshold being crossed
at this time comes from the Netherlands and New Zealand. In
the Netherlands, Zagwijn
14
has shown that there was a complete
© Nature Publishing Group1984
~N~ATU~R~E~V~O~L~
.~
30
~
7
~
1
~
6~F~E~BR~U~A~R~Y~19~84~
____________
----LETTERSTONATURE-------------------------------------6==13
Fig.S Carbon isotope records for Site 552A
cores
7-12
(top) and Pacific core V28-179
(bottom). Both are plotted on the same l3C
scale and represent estimates of the l3C con-
tents of intermediate depth (2,311
m)
water
in
the North Atlantic (above) and of deep
(4,509
m)
water
in
the equatorial Pacific
(below). Today there
is
an observed differ-
ence of a little over
1%
which reflects the fact
that ocean deep waters are ventilated in the
North Atlantic.
• • 0
C?
1.'
0-
I
1.0
Z
0.'
0
III
a:
0.0
4(
CJ
.....
4(
~
-1
.0
W
0-1.1
-2
.0
--I-,.....,.--r-,...,.-r-.,.......-,-,.-,-,..--r-r-r-r-r-r-,-,.-,-,.-",,-.,.-,.--,.....,.--r-,...,.-,-.,...,.-,-,.-,-,..--.-,-,..-,.--,-+--
1A
1.'
1.'
1.7 1.' 1
.•
1.0
2.1
2.2~
2.a
1.4
2
.•
2.'
2.7 2.' I
.•
'.0
S.1 S.2 I .a
1.4
I .• S.'
overturn in vegetational type just after the Gauss-Matuyama
boundary; indeed he regards this as representing the Pliocene-
Pleistocene boundary. In New Zealand, Stipp and others
15
showed that there was a major drop in sea level between about
2.4 and 2.6 Myr; again, they correlated this event, which they
interpreted as the first glacio-eustatic fall in sea level, with the
Pliocene-Pleistocene boundary. We disagree with these
opinions, only because
we
consider it well established that the
Pliocene-Pleistocene boundary at the newly proposed stratotype
at Vrica
16
,
as
well
as at previous Calabrian sections, has a much
younger age of about 1.6 Myr (ref. 17).
If
the boundary were
selected on the basis of a significant climatic event rather than
on historical precedent,
it
might appropriately be re-positioned
so
as
to coincide with the climatic change about 2.4 Myr, but
we
do not recommend such an approach to the definition of
geological boundaries.
The data shown in Fig. 4 also indicate that there was consider-
able
climatiC
variability, somewhere on the globe, even before
2.5 Myr.
It
is
well
established that glaciation in Iceland occurred
as
early as 3.1 Myr (ref. 18) although recent work suggests that
glaciation did not reach sea level
in
Iceland until about 2.0 Myr
(ref. 19) and there
is
also evidence for glaciation in the Sierra
Nevada, California, at about the same time
20
• However,
palaeooceanographic conditions on Rockall Bank do not suggest
severe climates in North-west Europe. Occasional grains of
ice-rafted debris do occur, and the nannofossil floras certainly
show marked variation, but the scale of variation was much less
before 2.4 Myr than it was after the event at that time.
Figure 5 shows the carbon isotope records for DSDP 552A
and V28-179. The data are plotted on the same isotope scale
and the same species-dependent correction factors have been
made to both. The persistent isotopic difference between the
two records
of
over
1%
indicates a continuing situation in which
the water bathing the sea floor on Rockall was isotopically
heavier for
13C
than deep water in the Pacific, as
is
the case
today. This difference reflects the difference between newly-
formed, oxygen rich North Atlantic deep water (NADW) and
older, oxygen-depleted deep Pacific water
21
,22.
We (and also
Blanc
et
alP)
disagree with Keigwin's deduction
24
that the
formation of NADW only began with the closing of the Panama
Straits at 3.0 Myr, although this event may well have affected
the contribution that this water mass made to the deep water
of the Caribbean deep waters that
we
re the topic of Keigwin's
stud
y
2
4.
In the upper part of the section, it
is
clear that DSDP
Site 552A preserves the whole of the standard oxygen isotope
AGE,
MILLION
YEARS
record of the past million years. Although the fine structure of
many stages
is
obscure, every single stage
is
present. By com-
parison, we note that the stratigraphically longest piston core
from the North Atlantic, core Kane 708-7 (ref. 9), extends only
to Stage 17 at about 0.6 Myr. For the earlier record, only DSDP
Site 116, recently re-investigated from the palaeoclimatic point
of view
25
, was previously available. Site 116 was recovered
by
conventional rotary drilling and the sediment was so disturbed
that glacial and non-glacial materials are intermixed, precluding
any high-resolution studies being made.
It
is
to be hoped that
the new technique of
13C
analysis in the organic carbon portion
of the sediment, pioneered in Site 116 (ref. 25), will be applied
to Site 552A. We believe that the material recovered in Site
552A
is
ideal for detailed studies of climatic evolution in the
ocean adjacent to the north-west European region where the
Plio-Pleistocene continental record
is
best known.
N.J.S. thanks NERC for support and Mike Tabecki for labora-
tory assistance. D.V.K. tanks NSF for support, Lamont-
Doherty Geological Observatory contribution no. 3552.
Received 22 August; accepted
31
October 1983.
1.
Berggren,
A.
A.
Init. Rep.
DSDP
n,
953-963 (1972).
2.
Backman, 1. Stockh. Contr. Geot.'31, 113-137 (1979
).
3.
Prell, W. L. et
aL
Init.
Rep
.
DSDP
68,
5-6
(1982).
4. Shackleton, N. J.
4<
Opdyke, N. D.
Quat
.
Res
. 3, 39-55 (1973).
5. Shackleton, N. J.
CoiL
into
Cent
Mtn. Reck. Sc
ient
119,203-210
(1974
).
6. Graham, D. W., Corliss,
B.
H., Bender, M. L.
4<
Keigwin, L.
D.
Jr
Mar
. Micropalaeont
6, 483-497 (1981).
7. Duplessy, J.-C.
et
aL
Quat
Res. (in the press).
8. Goree, W.
S.
4<
Fuller,
M.
Rev
. Geophys. Space Phys. 14, 591-608 (1976
).
9. Ruddiman,
W.
F.
4<
McIntyre, A. Mem. GeoL
Sac
. Am. 145, 111-146 (1976).
10. Imbrie, J.
et
aC
in
Milankovich
and
Climate (eds Berger, A. L. et aLl (Reidel, Dordrecht,
in
the press).
11. Shackleton,
N.l
.
4<
Opdyke, N. D. Nature 170, 216-219 (1977).
12. Berggren, W. A., Kent,
D.
V. & van Couvering, J. Geol. Soc. Lond. Spec. PubL (ed. Snelling,
N.
L.) (in the press).
13.
Backman, J. A. & Shackleton,
N.
J.
Mar
.
Mi
cropalaeont.
8,141-170
(1983).
14. Zagwijn, W.
Boreas 3, 75-97 (1974).
15
. Stipp, J. J., Chappell, J. H. A. & McDougall,
I.
Am.
J.
Sci.
265, 462-474 (1967).
16. Selli,
R.
et
aL
Giom. GeoL 41, 181-204 (1977
).
17. Backman, J., Shackleton,
N.
J.
&
Tau.e,
L. Nature 304, 156-158 (1983).
18. McDougall,
I.
& Wensink, H. Earth planet. Sci Lerr.
1,232-236
(1966).
19. Albert.son, K. J.
N6ttUrufraedingurinn '48, 1-8 (1978).
20. Curry,
R.
R. Science 154, 1121-1142 (1966).
21. Duplessy, J.-C. thesis, Univ. Paris VI (1972).
22
. Kroopnick, P., Deuser,
W.
G. & Craig, H.
J.
gtophys.
Res
. 75, 7668-7671 (1970).
23. Blanc, P. L., Rabussier,
D.
, Vergnaud-Grazzini, C. & Duplessy, J
.-
c.
Nature
%83
, 553-555
(1980).
24. Keigwin,
L.
D.
Science 117, 350-353 (1982).
25
. Blanc, P. L., Fontugne, M.
R.
& Duplessy, J .•
c.
Palaeogeogr. Palatoclimatol. Palae""coL
41,211-224
(1983).