Quaternary Climates and Sea Levels of the U.S. Atlantic Coastal Plain

Abstract

Uranium-series dating of corals from marine deposits of the U.S. Atlantic Coastal Plain coupled with paleoclimatic reconstructions
based on ostracode (marine) and pollen (continent) data document at least five relatively warm intervals during the last 500,000
years. On the basis of multiple paleoenvironmental criteria, we determined relative sea level positions during the warm intervals,
relative to present mean sea level, were 7 ± 5 meters at 188,000 years ago, 7.5 ± 1.5 meters at 120,000 years ago, 6.5 ± 3.5
meters at 94,000 years ago, and 7 ± 3 meters at 72,000 years ago. The composite sea level chronology for the Atlantic Coastal
Plain is inconsistent with independent estimates of eustatic sea level positions during interglacial intervals of the last
200,000 years. Hydroisostatic adjustment from glacial-interglacial sea level fluctuations, lithospheric flexure, and isostatic
uplift from sediment unloading due to erosion provide possible mechanisms to account for the discrepancies. Alternatively,
current eustatic sea level estimates for the middle and late Quaternary may require revision.

Full text

16
January
1981,
Volume
211,
Number
4479
SCI
E:
NCE
Regional
Geology
and
Coral
Ages
Quaternary
Climates
and
Sea
Levels
of
the
U.S.
Atlantic
Coastal
Plain
Thomas
M.
Cronin,
Barney
J.
Szabo,
Thomas
A.
Ager
Joseph
E.
Hazel,
James
P.
Owens
Recent
Quaternary
paleoclimate
stud-
ies
have
focused
mainly
on
isotopic
(1),
paleontologic
(2-4),
and
lithologic
(5)
data
from
deep-sea
cores
and
isotopic
studies
of
speleothems
(6,
7).
These
in-
vestigations
yielded
quantitative
data
for
understanding
patterns
of
Quaternary
climatic
variation
culminating
in
a
global
climatic
reconstruction
at
glacial
maxi-
mum
18,000
years
ago
(8)
and
a
demon-
margins
have
been
neglected
in
com-
bined
sea
level
and
paleoclimatic
stud-
ies.
We
selected
the
emerged
southern
Atlantic
Coastal
Plain
of
Virginia
and
North
and
South
Carolina
(Fig.
1)
to
in-
vestigate
middle
and
late
Pleistocene
(17)
climates
and
relative
sea
levels
because:
(i)
extensive
surface
and
shallow
sub-
surface
mapping
provides
a
lithostra-
tigraphic
framework
(18);
(ii)
relatively
Summary.
Uranium-series
dating
of
corals
from
marine
deposits
of
the
U.S.
Atlantic
Coastal
Plain
coupled
with
paleoclimatic
reconstructions
based
on
ostracode
(ma-
rine)
and
pollen
(continent)
data
document
at
least
five
relatively
warm
intervals
dur-
ing
the
last
500,000
years.
On
the
basis
of
multiple
paleoenvironmental
criteria,
we
determined
relative
sea
level
positions
during
the
warm
intervals,
relative
to
present
mean
sea
level,
were
7
+
5
meters
at
188,000
years
ago,
7.5
+
1.5
meters
at
120,000
years
ago,
6.5
+
3.5
meters
at
94,000
years
ago,
and
7
+
3
meters
at
72,000
years
ago.
The
composite
sea
level
chronology
for
the
Atlantic
Coastal
Plain
is
inconsistent
with
independent
estimates
of
eustatic
sea
level
positions
during
interglacial
intervals
of
the
last
200,000
years.
Hydroisostatic
adjustment
from
glacial-
interglacial
sea
level
fluctuations,
lithospheric
flexure,
and
isostatic
uplift
from
sediment
unloading
due
to
erosion
provide
possible
mechanisms
to
account
for
the
discrepancies.
Alternatively,
current
eustatic
sea
level
estimates
for
the
middle
and
late
Quaternary
may
require
revision.
strated
statistical
association
linking
changes
in
earth
orbital
parameters
with
Quaternary
interglacial-glacial
transi-
tions
(9).
The
magnitude
and
age
of
re-
sulting
glacio-eustatic
sea
level
fluctua-
tions
have
been
estimated
from
integra-
tion
of
uranium-disequilibrium-series
dating
of
coral
terraces
(10-12)
and
fluc-
tuations
in
oxygen
isotope
(180)
values
in
deep-sea
sediments.
These
fluctuations
have
been
attributed
to
changes
in
ocean
volume
(13,
14).
Yet
until
recently
(15,
16),
continental
SCIENCE,
VOL.
211,
16
JANUARY
1981
accurate
sea
level
indicators
are
avail-
able
(19);
(iii)
paleo-oceanography
can
be
reconstructed
from
ostracode
zoogeo-
graphy;
(iv)
pollen
preserved
in
marine
deposits
yields
climatic
data
for
the
adja-
cent
continent;
(v)
corals
suitable
for
uranium-series
dating
are
preserved;
and
(vi)
results
can
be
compared
with
pa-
leoclimatic
reconstructions
of
the
adja-
cent
open
ocean
North
Atlantic
(20)
and
can
be
related
to
continental
ice
sheet
growth
and
decline
in
the
Northern
Hemisphere.
Pleistocene
Coastal
Plain
geologic
his-
tory
records
marine
transgressive-re-
.gressive
cycles
in
sediments
deposited
in
brackish
water
and
marine
environments
(21,
22).
The
landward
limit
of
each
se-
quence
is
marked
by
an
erosional
scarp,
brackish
water
deposits,
barrier
islands,
and
beach
and
other
shoreline
deposits.
Seaward
of
these
paleo
shorelines
are
level
to
gently
inclined
(10
to
20)
plains
commonly
referred
to
as
terraces.
Strati-
graphic
unconformities
mark
periods
of
emergence
between
successive
high
stands
of
sea
level.
Early attempts
to
correlate
transgressions
solely
by
shore-
line
elevation
(23)
assumed
complete
tec-
tonic
stability
and
met
with
criticism
be-
cause
some
geomorphic
features
appear
to
be
tectonically
warped
(18,
24).
Lithologic
sections
of
12
studied
local-
ities
are
shown
in
Fig.
2;
locality
and
pa-
leontologic
data
for
each
section
are
giv-
en
in
Table
1.
Except
for
material
from
locality
6,
our
data
come
from
areas
where
lithostratigraphic
relationships
among
transgressive
sequences
have
been
studied.
However,
stratigraphic
nomenclature
is
in
a
state
of
confusion
in
parts
of
the
Atlantic
Coastal
Plain
(18,
25)
mainly
because
of
a
proliferation
of
local
morpho-,
litho-,
and
biostratigraph-
ic
terms,
many
of
which
have
been
vaguely
or
improperly
defined
and
sub-
sequently
misused
or
misinterpreted.
This
nomenclature
problem
is
especially
acute
with
regard
to
"terrace
forma-
tions"
(21,
23)
that
were
originally
named
for
a
geomorphic
surface,
not
a
lithologic
unit.
We
prefer
not
to
recon-
cile
stratigraphic
problems
nor
to
perpet-
uate
the
use
of
convenient
and
well-
known,
but
ambiguous
or
improper,
stratigraphic
names.
In
this
article,
we
do
not
use
terrace
formation
terms,
but
only
clearly
delineated
lithostratigraphic
terms.
With
the
exception
of
deposits
at
lo-
cality
8C
(Table
1),
which
are
early
Pleistocene,
all
sampled
units
are
middle
and
late
Pleistocene
in
age
(18,
22)
as
judged
by
field
relations,
several
lines
of
biostratigraphy
(22,
26),
and
preliminary
T.
M.
Cronin,
T.
A.
Ager,
J.
E.
Hazel,
and
J.
P.
Owens
are
members
of
the
U.S.
Geological
Survey
in
Reston,
Virginia
22092,
and
B.
J.
Szabo
in
Den-
ver,
Colorado
80225.
0036-8075/81/0116-0233$02.00/0
Copyright
0
1981
AAAS
233
on June 20, 2014www.sciencemag.orgDownloaded from on June 20, 2014www.sciencemag.orgDownloaded from on June 20, 2014www.sciencemag.orgDownloaded from on June 20, 2014www.sciencemag.orgDownloaded from on June 20, 2014www.sciencemag.orgDownloaded from on June 20, 2014www.sciencemag.orgDownloaded from on June 20, 2014www.sciencemag.orgDownloaded from on June 20, 2014www.sciencemag.orgDownloaded from

magnetostratigraphy
(27).
To
improve
dating
of
these
units,
we
relied
on
urani-
um-disequilibrium-series
dating
on
fossil
corals.
This
technique
has
been
widely
used
for
dating
Pacific
and
Caribbean
coral
reef
terraces
(11,
12)
and
corals
and
oolites
from
Florida
(28).
Our
dated
cor-
als
were
preserved
in
clastic
sediments
of
inner
sublittoral
depositional
environ-
ments,
not
in
limestone
reefs,
and
hence
we
use
corals
solely
to
date
trans-
gressions,
not
to
estimate
the
elevations
of
paleo
sea
levels.
Coral
samples
were
ultrasonically
cleaned
in
water,
ground
to
fine
powder,
and
ashed
at
900°C
for
about
6
hours.
Uranium
and
thorium
percentages
and
the
activity
ratios
of
234U
to
238U
and
2317h
to
234U
were
determined
by
alpha-
spectrometry
measurements
from
a
com-
bined
spike
solution
of
236U,
229f,
and
228Th
(29).
Analytical
results
and
coral
ages
are
presented
in
Table
2.
Fourteen
uranium-series
dates
yielded
the
follow-
ing
clusters
of
coral
ages:
(i)
more
than
400,000
years
old
(localities
7
and
8),
(ii)
about
188,000
years
old
(localities
1
and
6);
about
120,000
years
old
(locality
11);
about
94,000
years
old
(localities
9,
10,
and
12);
and
(v)
about
72,000
years
old
(localities
2
to
5).
Ostracodes
and
Pollen:
Paleoclimatic
Tools
Oceanic
paleoclimate
studies
rely
on
planktic
floras
and
faunas
to
reconstruct
sea-surface
temperatures
(2-4,
30,
31);
continental
studies
emphasize
pollen
(32).
Ostracodes
are
benthic,
shelled
Crustacea,
which
traditionally
are
used
for
biostratigraphy
but
more
frequently
are
being
used
in
freshwater
(33)
and
ma-
rine
continental
shelf
(34-36)
paleocli-
matic
reconstructions.
Many
species
have
zoogeographic
distributions
limited
by
the
water
temperatures
necessary
for
their
reproduction
and
survival
(37,
38).
North
Atlantic
ostracode
faunas
migrat-
ed
south
several
hundred
kilometers
in
response
to
climatic
cooling
associated
with
the
Wisconsin
glaciation
(34,
36).
This
migration
demonstrates
the
temper-
ature
sensitivity
of
marine
ostracodes
and
underscores
their
paleoclimatic
po-
tential,
particularly
along
continental
margins
where
they
are
widely
preserved
as
fossils
and
where
planktic
groups
are
rarely
preserved.
Five
faunal
provinces
have
been
defined
in
the
western
North
Atlantic.
Four
of
these
provinces
corre-
spond
to
modern
frigid,
subfrigid,
mild
temperate,
and
subtropical
marine
cli-
matic
zones
(35,
37,
38).
A
warm
temper-
ate
zone,
in
evidence
in
Virginia
during
the
late
Pleistocene
(35)
(localities
2,
3,
and
5
in
our
study),
is
missing
today
due
to
strong
isothermal
convergence
at
1
1
Sand
ECIay
LI1lit
Ra
ar
oC
ross
LLocality
oo
sOtracode
1
Pollen
s
u
Coral
uranium-
Estimated
rango
of
sample
sample
sorle
sample
maximum
sea
lvois
Fig.
1
(left).
Locations
of
Coastal
Plain
Quaternary
outcrops
yielding
corals
for
uranium-series
dating
and
ostracodes
and
pollen
for
paleoclimate
study.
Fig.
2
(right).
Lithological
sec-
tions,
elevation,
estimated
range
of
maximum
sea
levels,
and
location
of
ostracode
pollen
and
coral
samples
for
12
studied
localities.
Section
for
locality
5
had
to
be
inferred
from
dredged
material
because
the
pit
was
filled
with
water,
and
the
coral
from
this
locality
was
not
in
place.
Predominant
molluskan
genera
are
given
for
some
beds.
SCIENCE,
VOL.
211
234

-
Ei.l.
,
l,
,8
8
o
3s
;s
4
o
°
g
C
|
3,
8
8
8
o
8
A
;
"
f
4
^
g8
o
8
8
|
g~~~~~~~~~~~~.+I
q
%
%0
N-
+1~~
1
+'
+1 +1
+i
|1
E
d
d
W
ir1
iitd
d^t
i
}|
|t0tt
dti
l-d
c
<
S
t}G
8
i
i|]
i[]tSgg[i.,g
Sg
SS
aig[gj
l~~~~~~~~~~~~~~~~~~~~~W
2
.
E
g
XE
tz
t
tX
g
t
y'^Y
t
R
t
;g
Y
i
t
:
33~~~~~~~~~~~~~~~~~~4
.
E
AF~~~~~~
0
z
z
1 8
°
z
>
*
^
r; r;
n^
W
X
#A
co
A
.
4
>
ci
Go
>
Ai
z
>
Z
W
C-00
.
8
6:
.
G02
GO
E
9
0
z
0
4+
4+
-1+
W-4
tn
a
00
00
C4
JANUARY
1981
235
16

Western
North
Atlantic
zoogeography
and
cUlmates
Bottom
Continental
margin
Surface
Open
ocean
temperature
(OC)
Sublittoral
Marine
CC)
Planktic
zoogeographic
climatic
zone
foraminiferal
Winter
Summer
province
Winter
Summer
assemblages
Arctic
Frigid
l
[___--0___-H
II
~~~~~~~~~Polar
I
Labrador
Subfrigid
0
10
-
__
Subpolar
15-
Nova
Scotian
Cold
temperate
5-
21
20-
5-
25
Virginian
Mild
temnperate
TransHtional
10-
=
-l
-
-
_
15-
26
Warm
eemDat
15
-
Carolinian
Subtropical
Subtropical
20°-
30-
22-
29-
"Caribbean"
Tropical
Tropical
Fig.
3.
Modern
zoogeographic
provinces
and
marine
climate
zones
of
the
western
North
Atlan-
tic.
Sublittoral
zoogeography
and
climate
zones
are
adopted
from
Hazel
(38).
Planktic
forami-
niferal
open-ocean
data
are
summarized
from
Imbrie
and
Kipp
(2)
and
Kipp
(1976)
(31);
bounda-
ries
between
assemblages
are
approximate.
Winter
(February)
and
summer
(August)
mean
sea-
surface
temperatures
of
the
open
ocean
are
taken
from
(62).
These
temperatures
were
measured
in
the
southwestern
North
Atlantic
at
the
following
latitudes
and
longitudes:
500N,
50°W;
42°N,
60°W;
35°N,
70°W;
and
26°N,
75°W.
This
figure
shows
only
a
rough
correlation
between
open-
ocean
and
sublittoral
climatic
zones.
Warm
temperate
zone
does
not
exist
today
because
of
strong
isothermal
convergence
on
the
shelf
at
Cape
Hatteras.
Cape
Hatteras.
The
known
modern
dis-
tribution
of
ostracodes
in
these
prov-
inces
serves
as
the
ecological
data
base
for
our
study
(35,
37,
38).
In
Fig.
3,
we
relate
this
sublittoral
scheme
to
the
open
ocean
classification
(2-4,
30,
31),
which
is
based
on
modern
distributions
of
North
Atlantic
planktic
foraminifers.
Paleotemperature
estimates
of
all
dated
ostracode
assemblages
consistent-
ly
indicate
oceanic
bottom-water
tem-
peratures
as
warm
or
warmer
than
pres-
ent
interglacial
temperatures
offshore
at
the
same
latitude
(Fig.
3)
(35).
The
local-
ities
yielding
coral
ages
older
than
400,000
years
(localities
8,
A
and
B),
which
are
south
of
Cape
Hatteras
in
a
modern
subtropical
climatic
zone,
also
yielded
samples
with
subtropical
ostra-
code
assemblages.
Our
localities
dated
at
188,000
years
(localities
1
and
6)
are
north
of
Cape
Hatteras
in
a
modern
mild
tem-
perate
zone,
and
they
contain
marine
subtropical
fossil
assemblages.
We
inter-
pret
this
northward
migration
of
thermo-
philic
ostracode
species
as
a
reflection
of
incursions
of
warm
water,
probably
from
the
Gulf
Stream,
into
the
present
Pam-
lico
Sound
and
Chesapeake
Bay
regions
during
an
interglacial
interval.
Rare
tem-
perate
species
(39)
at
locality
6
may
sig-
236
nify
a
minor
cooling
in
this
region,
but
this
interpretation
needs
further
con-
firmation.
In
South
Carolina,
our
locality
dated
at
120,000
years
(locality
11)
and
the
three
localities
with
ages
clustering
at
94,000
years
(localities
9,
10,
and
12)
contain
subtropical
assemblages
signify-
ing
warm
interglacial
climates.
Several
undated
localities
near
Charleston,
which
we
interpret
to
be
correlative
with
the
transgression
at
94,000
years,
con-
tain
ostracode
species
known
only
from
the
Gulf
of
Mexico
(35).
Their
occur-
rence
in
the
Atlantic
Coastal
Plain
im-
plies
faunal
interchange
with
the
Gulf
and
winter
water
temperatures
slightly
higher
than
those
off
South
Carolina
today.
In
southern
Virginia
and
northern
North
Carolina,
three
of
the
four
local-
ities
with
ages
averaging
72,000
years
(localities
2,
3,
and
5)
contain
warm
tem-
perate
ostracode
assemblages,
indicative
of
temperatures
higher
in
winter
and
comparable
in
summer
to
present
tem-
peratures,
which
represent
mild
temper-
ate
conditions
(Table
1
and
Fig.
3).
Fossil
pollen
and
spores
from
samples
of
unoxidized
clay,
clayey
sand,
and
peat
collected
from
the
coral-
and
ostra-
code-bearing
deposits
provide
insight
in-
to
contemporaneous
onshore
vegetation
environments
and
climates.
The
pollen
data
provide
some
of the
first
documen-
tation
for
pre-Sangamonian
(earlier
than
isotope
stage
5)
Pleistocene
interglacial
vegetation
of
the
southern
Atlantic
Coastal
Plain.
Previous
investigators
of
pollen-bearing
deposits
of
the
region
(40-
42)
presumed
these
deposits
were
San-
gamonian
in
age.
Two
such
deposits
(the
Canepatch
and
Flanner
Beach
Forma-
tions)
now
appear
to
be
middle
Pleisto-
cene
in
age.
A
variety
of
environments,
including
salt
marsh,
coastal
and
fluvial
swamps,
and
coastal
plain
forests,
are
represented
in
the
pollen
assemblages.
With
two
ex-
ceptions
(localities
6
and
8A),
all
samples
yielded
interglacial
pollen
assemblages
similar
to
late
Holocene
pollen
assem-
blages
from
the
same
region.
The
inter-
glacial
fossil
floras
include
primarily
oak
(Quercus),
pine
(Pinus),
and
hickory
(Carya)
pollen,
and
usually
smaller
per-
centages
of
taxa
such
as
sweet
gum
(Liq-
uidambar),
black
gum
(Nyssa),
birch
(Betula),
and
others.
This
pollen
assem-
blage
suggests
a
warm
and
humid
tem-
perate
to
subtropical
climate
similar
to
that
existing
today
in
the
Atlantic
Coast-
al
Plain
of
South
Carolina,
North
Caroli-
na,
and
Virginia
(43).
Further,
the
pollen
data
support
the
interpretation
that
past
high
stands
of
sea
level
for
the
southern
Atlantic
Coastal
Plain
coincided
with
warm,
interglacial
terrestrial
climates.
The
only
exceptions
to
the
pattern
of
interglacial
pollen
assemblages
are
sam-
ples
from
localities
6
and
8A
which,
in
addition
to
pollen
of
oak,
pine,
hickory,
and
other
common
deciduous
trees,
con-
tain
small
amounts
(generally
less
than
2
percent
each)
of
spruce
(Picea)
and
hem-
lock
(Tsuga)
pollen.
Studies
of
Wiscon-
sinan
age
deposits
in
the
study
region
(40,
44-46)
indicate
that
boreal
elements
such
as
spruce
(probably
predominantly
white
spruce,
Picea
glauca),
hemlock,
and
jack
pine
(Pinus
banksiana)
migrat-
ed
far
south
of
their
present
distributions
and
became
important
elements
of
the
regional
vegetation
of
the
southern
At-
lantic
Coastal
Plain.
Red
spruce
(Picea
rubens)
and
hemlock
(Tsuga
canadensis
and
T.
caroliniana)
now
grow
in
parts
of
the
central
and
southern
Appalachians
(47).
The
presence
of
small
amounts
of
spruce
and
hemlock
pollen
with
the
pre-
dominant
oak-pine-hickory
assemblage
from
localities
6
and
8A
indicates
tem-
perate
climates,
cooler
than
those
of
today
but
warmer
than
a
full
glacial
cli-
mate.
These
data
and
the
occurrence
of
temperate
ostracode
species
at
locality
6
suggest
a
transitional
climate,
inter-
mediate
between
glacial
and
intergla-
cial.
SCIENCE,
VOL.
211
80-
z
°0
60-
-J
40-
20-
0

Relative
Paleo
Sea
Level
We
estimated
the
ranges
of
maximum
relative
paleo
sea
levels
above
mean
sea
level
(ASL)
for
each
locality
using
vari-
ous
shoreline
criteria
(Fig.
2).
Sea
level
ranges
were
then
determined
for
each
of
.
the
coral
age
clusters
at
440,000,
0-
50-
a
,,
1
00
-
100
0
-.
150-
200
Southem
Atlantic
Coastal
Plain,
composite
0
aa
I;-
0
I
I
I
I+
I
I
I
New
Guinea
00c
I
a
I
I
-IV
-4
VI
q
vI
>=VVI
>Vlla
188,000,
120,000,
94,000,
and
72;000
years
ago
from
sea
level
ranges
of
each
locality
in
each
cluster.
About
440,000
years
ago
(localities
8,
A
and
B)
in
north-
eastern
South
Carolina,
maximum
sea
level
was
between
10
and
25
meters
ASL
as
judged
by
faunal
paleodepth
estimates
(19)
and
barrier
island
deposits.
Pa-
Bermuda
Barbados
V
I
10
0
>
Venftno
R
inelndevot
Kel
alH
KRelindevu
~)Aberdare
leontologic
evidence,
including
the
pres-
ence
of
Callianassa
burrows,
nearshore
faunal
assemblages
from
southeastern
Virginia
(locality
1),
and
Mulinia
-rich
de-
posits
in
central
North
Carolina
(locality
6),
indicate
a
range
of
sea
levels
between
2
and
12
m
ASL.
Locality
11,
dated
at
120,000
years
ago,
contains
a
fauna
in-
Generalzed
oxygen
botope
curve
1
9)
3
5
7
Marine
paleoclimates
South-
Eastern
North
eastern
South
Atlatis
Vinlrg
Carolna
50°N
MItD
-Zfi
suTRap-A
I
RNSrNAL
_W_
_&_
WARM
TEbtPERATE
SUBTROPKCAL.
SUBFIPUAL
ALTERNATING
POLAR
AND
SUBPOLAR
SUBPOLAR
TRANSMONAL
SUBPOLA
TRANS
ITONAL
COLD
SUB
Llg
Fig.
4.
U.S.
Atlantic
Coastal
Plain
sea
levels
and
marine
paleoclimates
compared
with
New
Guinea
(11),
Bermuda
(7),
and
Barbados
(14)
sea
level
chronologies
and
North
Atlantic
(4)
paleoclimatic
history
at
50°N.
Fairbanks
and
Matthews
(14)
assigned
an
arbitrary
error
value
of
+20
m
on
all
Barbados
sea
levels
except
Rendevous
Hill.
For
Bermuda,
not
all
speleothems
of
Harmon
et
al.
(7)
are
shown.
The
U.S.
Coastal
Plain
data
show
the
estimated
range
of
maximum
sea
levels
for
four
uranium-series
age
clusters
and
analytical
error
for
dates.
This
is
a
composite
figure
represent-
ing
several
geographic
regions
that
may
have
different
tectonic
histories.
Under
the
marine
paleoclimates
column,
the
range
of
uranium-series
ages
for
each
cluster
is
given
by
solid
lines;
the
analytical
errors
by
dashed
lines.
Abbreviation:
MSL,
mean
sea
level.
Table
2.
Analytical
data
and
uranium-series
ages
of
fossil
corals
from
deposits
of
the
southeastern
Atlantic
Coastal
Plain.
Lo-
Coral
Cal-
Uranium
Activity
ratios
Age
cal-
ge-
cite
(ppm)
ity
nus*
(%7O)
(ppm)
2UfU
23h32Th
2W
u34
(U03
years)
-
72,000-year-old
deposits
3
A
<
3
2.53
+
0.02
1.09
±
0.02
15.0
0.436
±
0.017
62
±
4t4
2
A
<
3
2.64
+
0.03
1.10
+
0.02
25.0
0.500
±
0.020
74
±
4t
5
A
<
3
2.98
+
0.04
1.10
+
0.02
39.0
0.487
±
0.019
72
±
4t
4
A
<
3
2.44
+
0.05
1.11
±
0.02
24.0
0.524
±
0.021
79
±
5t
94,000-year-old
deposits
10
A
<
3
3.42
±
0.07
6.11
±
0.12
2.0
0.594
+
0.024
96
±
6t§
12
A
<
3
2.79
±
0.04
1.09
±
0.02
2.0
0.581
+
0.023
93
±
6t§
9
A
<
3
2.66
±
0.04
1.07
±
0.02
7.1
0.586
±
0.023
94
±
6t§
-
120,000-year-old
deposits
11
A
<
3
2.86
±
0.04
1.09
+
0.02
13.0
0.677
±
0.020
120
±
6t
-
188,000-year-old
deposits
I
A
<
3
2.27
±
0.04
1.06
±
0.02
11.0
0.832
±
0.033
187
±
20t
6
A
<
3
2.69
±
0.04
1.07
±
0.02
80.0
0.837
±
0.025
189
±
15t
>
400,000-year-old
deposits
8A
S
<
3
2.50
±
0.02
1.04
±
0.01
14.0
1.01
±
0.03
560
+
x,
-190t
(440
±
140)11
8B
S
<
3
2.99
±
0.04
1.04
±
0.01
160.0
0.982
±
0.029
376
+
201,
-68t
(440
+
140)11
8C
S
<
3
2.62
±
0.03
1.017
±
0.010
500.0
1.017
±
0.030
>
424t;
740
+
360,
-21011
7
S
<
3
2.70
±
0.04
1.02
±
0.02 55.0
1.04
±
0.04
>
500t
*Coral
genera:
A
is
Astrangia
sp.
and
S
is
Septastrea
sp.
tThe
uhlwh
ages
were
calculated
from
half-lives
of
Luh
and
2"U
of
75,200
and
244,000
years,
respective-
ly.
tThe
3'Pa
age
is
64,000
+
8,000
years.
§The
21wh
age,
corrected
for
initial
23Ih
contamination
from
the
blh/V4U
versus
m2Th/34U
isochron
plot
(60),
is
94,000
+
6,000
years.
Irrhe
n4U
ages
were
calculated
from
the
average
n4U/3"U
ratio
in
Atlantic
Ocean
waters
of
1.14
±
0.02
(61)
by
(SUU/rnU)
sample
-
1]
=
in
1.14-1I
16
JANUARY
1981
237

dicating
that
paleo
sea
levels
were
be-
tween
6
and
9
m
ASL.
At
94,000
years
ago,
sea
levels
in
South
Carolina
be-
tween
3
and
10
m
ASL
are
in
evidence
from
deposits
with
high
angle
cross-bed-
ding
interpreted
as
tidal
delta
in
origin,
from
beach
and
dune
sands,
and
from
faunal
assemblages
(localities
9,
10,
and
12).
In
the
Virginia-North
Carolina
bor-
der
area,
at
72,000
years
ago,
relative
sea
levels
between
4
and
10
m
ASL
are
in-
dicated
by
in
situ
oysters
(Crassostrea
virginica),
cross-bedding,
beach
sands,
and
nearshore
faunal
assemblages.
We
emphasize
that
these
paleo
sea
level
po-
sitions
are
estimates;
their
significance
lies
in
the
fact
that
each
is
above
present
sea
level.
As
the
first
uranium-series
study
of
this
region,
our
composite
chronology
must
be
considered
a
first
approxima-
tion.
Nevertheless,
the
uranium-series
ages
are
augmented
by
detailed
paleocli-
matic
data,
biostratigraphy,
and
litho-
stratigraphy,
and
we
propose
(Fig.
4)
some
correlations
to
deep-sea
isotope
stages
and
to
coral
terraces
in
Barbados,
which
have
been
used
to
date
isotope
stages
(14).
The
740,000-year
age
on
cor-
al
from
the
Waccamaw
Formation
(local-
ity
8C)
approaches
the
limit
of
the
dat-
ing
technique,
and
we
view
it
with
cau-
tion.
However,
it
is
generally
consistent
with
biostratigraphic
and
magnetostrat-
igraphic
data
indicating
an
early
Pleisto-
cene
age
(22,
27).
We
correlate
the
trans-
gression
of
440,000
years
ago
in
South
Carolina
(localities
8,
A
and
B),
and
per-
haps
the
500,000-year
date
from
North
Carolina,
with
isotope
stages
9,
11,
or
possibly
13.
Isotope
stage
7,
dated
at
180,000
to
220,000
years
ago,
from
the
Aberdare,
Kingsland,
and
Kendall
Hill
Barbados
terraces,
is
probably
repre-
sented
by
our
188,000-year-old
unit.
The
anticipated
predominance
of
ages
near
120,000
years
(the
peak
interglacial
stage
5e
at
125,000
±
6,000
years
ago,
repre-
sented
by
the
Rendevous
Hill
Barbados
terrace)
is
not
in
evidence-only
one
120,000-year
age
may
represent
this
peri-
od.
We
hesitate
to
rely
heavily
on
a
single
date.
The
Worthing
and
Ventnor
Barbados
terraces
date
late
isotope
stage
5a
at
82,000
+
5,000
years
and
stage
5c
at
105,000
+
5,000
years.
These
two
high
stands
of
sea
level
may
be
repre-
sented
by
our
cluster
of
four
dates
aver-
aging
72,000
+
5,000
years
and
three
dates
averaging
94,000
+
5,000
years,
re-
spectively.
It
is
unclear
whether
or
not
the
discrepancies
between
the
Atlantic
Coastal
Plain
and
the
Barbados
uranium-
series
dates
represent
real
age
dif-
ferences
of
several
thousand
years.
If
confirmed
as
real,
these
differences
might
signify
differential
glacio-isostatic
effects
on
observed
sea
levels
in
the
two
regions.
Paleoclimatic
Inferences
800
700
600
500
400
300
20°
80
0
700
600
50
0
40
0
Fig.
5.
Western
North
Atlantic
oceanographic
reconstruction
during
late
oxygen
isotope
st
5,
between
70,000
and
100,000
years
ago.
Position
of
paleo
Gulf
Stream
and
inferred
cyclc
storm
tracks
indicate
the
existence
of
a
strong
thermal
gradient
between
the
relatively
w;
North
Atlantic
and
the
North
American
Laurentide
ice
sheet
and
ice
shelves
[after
(20)].
margin
positions
are
speculative.
The
18°C
isotherm
for
isotope
stage
5
to
4
transition
a
Ruddiman
and
McIntyre
(20).
238
Our
findings
have
a
bearing
upon
sev-
eral
distinct
but
interrelated
Quaternary
problems.
Laurentide
ice
sheet
(48)
and
oceanic
(20)
reconstructions
portray
North
Atlantic
conditions
during
ice
growth,
just
before
the
major
Wisconsin
glacial
interval
(isotope
stage
5
to
4
tran-
sition)
(Fig.
5),
as
having
sea-surface
temperatures
as
warm
as
or
slightly
cool-
300
er
than
modem
temperatures.
Rapid
growth
of
the
adjacent
North
American
ice
sheet
was
believed
to
have
been
caused
by
a
strong
thermal
gradient
be-
tween
ocean
and
ice
that
produced
cy-
clonic
storms
over
the
ice
sheet.
Our
cli-
matic
data
from
South
Carolina
and
the
Virginia-North
Carolina
border
area
for
late
stage
5
support
data
(20,
48,
49)
for
such
a
scenario
by
providing
direct
evi-
dence
for
a
relatively
warm
western
North
Atlantic
Ocean,
almost
surely
re-
2
flecting
the
Gulf
Stream
passing
along
20
the
southeastern
United
States
at
about
72,000
and
94,000
years
ago
(Fig.
5).
Northeastward
deflection
of
this
current,
as
exists
today,
would
produce
the
rela-
tively
warm
subpolar
mid-North
Atlan-
tic
that
is
in
evidence
from
the
deep-sea
data
(20,
49).
One
of
the
late
stage
5
Coastal
Plain
intervals
probably
also
cor-
relates
with
the
St.
Pierre
interstade
of
10
the
St.
Lawrence
Valley
and
the
Cape
Broughton
interstade
of
eastern
Baffin
Island
(48).
Further,
our
sea
level
high
stand
of
188,000
years
ago
appears
to
correlate
with
the
isotope
stage
7
to
6
age
transition
(49).
)mc
The
emerged
Coastal
Plain
is
the
land-
am
ward
part
of
the
subsiding
wedge
of
fter
Mesozoic
and
Cenozoic
sediments
that
constitute
the
Atlantic's
passive
conti-
SCIENCE,
VOL.
211

nental
margin.
Although
postrifting
sub-
sidence
rates
from
2
to
4
centimeters
per
103
years
characterize
the
geosynclinal
depocenters
of
the
adjacent
continental
slope
(50),
net
uplift
is
suspected
for
parts
of
the
emerged
Coastal
Plain
for
the
last
several
hundred
thousand
years.
We
compiled
a
composite
sea
level
dia-
gram
from
our
data
for
comparison
with
other
sea
level
chronologies
(Fig.
4)
to
show
the
combined
effects
of
glacio-eus-
tasy
and
regional
crustal
movements
on
our
relative
sea
level
record.
Clearly,
our
paleoclimatic
data
support
the
hypothe-
sis
that
there
is
a
major
glacio-eustatic
component
to
the
Coastal
Plain
record
because
high
stands
occurred
during
in-
tervals
of
demonstrably
warm
climates
both
onshore
and
offshore.
How
much
uplift
occurred
depends
on
how
close
es-
timates
of
eustatic
sea
levels
during
these
interglacial
intervals
are
compared
to
evidence
from
observed
sea
levels.
In-
dependent
estimates
for
middle
and
late
Pleistocene
interglacial
sea
levels,
based
on
uranium-series
dating
of
coral
ter-
races
and
oxygen
isotopic
studies,
call
for
a
eustatic
sea
level
of
+6
±
2
m
ASL
during
isotope
stage
5e
(about
125,000
years
ago)
(11,
14,
51).
Sea
level
during
other
middle
and
late
Pleistocene
warm
intervals
supposedly
remained
below
present
sea
level.
Recent
isotope
stage
7
estimates
are
-32
m
(220,000
years
ago),
-12
m
(200,000
years
ago),
and
-22
m
(180,000
years
ago)
(14).
However,
Shackleton
(52)
predicted
stage
7
sea
lev-
el
at
5
+
5
m,
which
is
remarkably
close
to
our
relative
sea
level.
Estimates
for
late
isotope
stage
5
are
-15
m
(11),
-16
m
(53),
and
-43
m
(14)
for
stage
Sc
(105,000
years
ago)
and
-13
m
(11),
-15
m
(53),
and
-45
m
(14)
for
stage
Sa
(82,000
years
ago).
If,
for
a
moment,
we
accept
the
range
of
eustatic
sea
level
estimates
of
-13
to
-45
m
for
late
isotope
stage
5,
and
-12
to
-32
m
for
stage
7,
and
also
our
corre-
lations
of
Atlantic
Coastal
Plain
trans-
gressions
with
the
isotope
stages
and
dated
coral
terraces
of
Barbados,
then
an
explanation
must
be
found
for
the
dis-
crepancies
in
contemporaneous
sea
level
positions.
One
possible
explanation
would
be
to
invoke
local
crustal
uplift
of
the
Atlantic
coast,
which
would
require
uplift
rates
of
roughly
0.20
to
0.45
milli-
meter
per
year
for
the
Charleston
and
southeastern
Virginia
regions
since
late
isotope
stage
5,
and
rates
of
about
0.10
to
0.20
mm
per
year
since
stage
7
for
the
region
from
southeastern
Virginia
to
northeastern
North
Carolina.
These
rates
of
uplift
appear
anomalously
high
for
a
relatively
stable
intraplate
coast,
especially
when
compared
with
rates
of
16
JANUARY
1981
0.20
mm
per
year
(54)
which
characterize
Barbados,
a
tectonically
active
(uplifted)
island.
Sediment
unloading
due
to
ero-
sion
may
have
caused
some
isostatic
up-
lift
but
probably
not
as
much
as
that
ob-
served.
Another
possible
mechanism
that
would
account
for
the
emerged
marine
deposits
on
the
Coastal
Plain
is
hydro-
isostasy-crustal
isostatic
adjustment
to
the
redistribution
of
mass
from
continen-
tal
ice
to
ocean
water
during
glacial-in-
terglacial
transitions.
Bloom
(55)
hypoth-
esized
that
this
redistribution
might
pro-
duce
an
amount
of
coastal
submergence
on
nonglaciated,
nonorogenic
continen-
tal
coasts
(such
as
the
Coastal
Plain)
pro-
portional
to
the
proximity
of
the
coast
to
deep
water.
Walcott
(56)
proposed
that,
rather
than
coastal
submergence,
coastal
uplift
occurred
as
an
increase
in
water
volume
depressed
ocean
basins
and
adja-
cent
continents
rose
from
redistribution
of
mantle
mass.
Clark
et
al.
(57)
ex-
panded
this
idea
and
suggested
that
all
coastal
regions
would
undergo
this
hy-
droisostatic
uplift
and
tilt.
The
earth's
rheological
response
to
the
last
deglacia-
tion
has,
however,
only
recently
been
discussed
in
detail
(58),
and
the
net
effect
of
multiple
hydroisostatic
deglacial
events
such
as
those
recorded
in
the
Coastal
Plain
would
be
difficult
to
pre-
dict.
This
mechanism
may
have
contrib-
uted
to
the
relatively
high
sea
levels
dur
ing
the
middle
and
late
Pleistocene
on
the
Coastal
Plain,
roughly
20
to
40
m
above
some
estimates
for
late
isotope
stage
5
and
stage
7
eustatic
sea
levels
(11,
53),
and
perhaps
also
for
the
relative
sea
levels
as
high
as
30
m
on
the
Coastal
Plain
during
the
early
Pleistocene
(22).
The
distinct
possibility
that
current
models
of
Quaternary,
eustatic
sea
level
fluctuations
might
need
revision
must
be
entertained
in
light
of
our
insufficient
un-
derstanding
of
Atlantic
continental
mar-
gin
tectonics
and
accumulating
evidence
on
sea
levels
that
seems
to
be
inconsist-
ent
with
these
models
(59).
Specifically,
estimates
of
glacio-eustatic
ice
volume
fluctuations
and
sea
level
high
stands
in-
ferred
from
stable
isotope
records
in
deep-sea
cores
may
be
no
more
accurate
than
those
estimates
of
high
stands
that
are
based
on
observed
paleo
sea
levels
on
continental
and
island
margins
and
that
have
been
corrected
for
neotectonic
vertical
movements
caused
by
geologic
factors
such
as
proximity
to
ice
sheets,
location
on
a
passive
plate
margin,
and
proximity
to
the
subsiding
depocenter
of
a
geosynclinal
trough.
This
latter
factor
is
pertinent
to
the
Atlantic
margin
be-
cause
a
region
like
Albemarle
Sound,
near
our
locality
5,
is
within
the
southern
part
of
the
subsiding
Baltimore
Canyon
trough,
while
eastern
South
Carolina
(lo-
calities
8
to
12)
is
several
hundred
ki-
lometers
from
the
Carolina
trough.
Dis-
tinct
trends
in
neotectonic
vertical
crus-
tal
movements
could
be
expected
in
these
two
regions
because
of
their
posi-
tions
with
respect
to
offshore
troughs-
specifically
subsidence
in
Albemarle
Sound
and
perhaps
lithospheric
flexural
uplift
in
eastern
South
Carolina.
But
our
knowledge
of
the
Quaternary
history
of
these
regions
is
still
insufficient
to
con-
firm
this.
Conclusion
Our
study
of
pre-Holocene
sea
levels
and
climates
on
a
nonglaciated,
nonoro-
genic
continental
coast
allows
us
to
pro-
pose
a
correlation
between
deep-sea
iso-
tope
stages
and
coral
terrace
chronolo-
gies
that
has
some
discrepancies
and
some
consistencies
with
current
eustatic
models.
We
believe
eustatic
sea
levels
for
stage
7
were
probably
near
present-
day
sea
level
and
that
-13
and
-15
m
are
more
realistic
estimates
for
eustatic
sea
level
during
stages
5a
and
5c
than
-45
and
-43
m.
In
general,
our
climatic
in-
ferences
match
North
Atlantic
deep-sea
data
on
the
timing
of
warm
climatic
inter-
vals
at
188,000,
120,000,
94,000,
and
72,000
years
ago.
We
also
conclude
that
there
is
a
primary
glacio-eustatic
com-
ponent
and
probably
a
secondary
neo-
tectonic
vertical
component
to
the
local
Coastal
Plain
sea
level
record
which
must
be
considered.
The
mechanisms
for
Quaternary
verticle
crustal
movements
probably
include
hydroisostasy
and
crustal
subsidence
and
uplift
caused
by
long-term
sediment
accumulation
in
dep-
ositional
troughs;
the
magnitude
of
crus-
tal
movement
probably
varies
along
the
entire
segment
of
the
coast
that
was
studied.
References
and
Notes
1.
C.
Emiliani,
J.
Geol.
63,
538
(1955);
W.
S.
Broecker
and
J.
van
Donk,
Rev.
Geophys.
Space
Phys.
8,
169
(1970);
N.
J.
Shackleton
and
N.
D.
Opdyke,
Quat.
Res.
(N.
Y.)
3,
39
(1973).
2.
J.
Imbrie
and
N.
G.
Kipp,
in
The
Late
Cenozoic
Glacial
Ages,
K.
K.
Turekian,
Ed.
(Yale
Univ.
Press,
New
Haven,
Conn.,
1971),
p.
71.
3.
C.
Sancetta,
J.
Imbrie,
N.
G.
Kipp,
A.
McIn-
tyre,
W.
F.
Ruddiman,
Quat.
Res.
(N.
Y.) 2,
363
(1972).
4.
W.
F.
Ruddiman
and
A.
McIntyre,
Geol.
Soc.
Am.
Mem.
145,
111
(1976).
5.
J.
E.
Damuth,
Geol.
Soc.
Am.
Bull.
88,
695
(1977);
W.
F.
Ruddiman,
ibid.,
p.
1813.
6.
P.
T.
Thomson,
H.
P.
Schwarcz,
D.
C.
Ford,
ibid.
87,
1738
(1976).
7.
R.
S.
Harmon,P.
T.
Thompson,
H.
P
Schwarcz,
D.
C.
Ford,
Quat.
Res.
(N.
Y.)
9,
54
(1978).
8.
CLIMAP
Project
Members,
Science
191,
1131
(1976).
9.
W.
S.
Broecker,
ibid.
151,
299
(1966);
J.
D.
Hays,
J.
Imbrie,
N.
J.
Shackleton,
ibid.
194,
1121
(1976).
239

10.
W.
S.
Broecker,
D.
L.
Thurber,
J.
Goddard,
T.
L.
Ku,
R.
K.
Matthews,
K.
J.
Mesolella,
ibid.
159,
297
(1968);
K.
J.
Mesolella,
R.
K.
Mat-
thews,
W.
S.
Broecker,
D.
L.
Thurber,
J.
Geol.
77,
250
(1969);
H.
H.
Veeh
and
J.
M.
A.
Chap-
pell,
Science
167,
862
(1970).
11.
A.
L.
Bloom,
W.
S.
Broecker,
J.
M.
A.
Chap-
pell,
R.
K.
Matthews,
K.
J.
Mesolella,
Quat.
Res.
(N.Y.)
4,
169
(1974).
12.
H. H.
Veeh,
J.
Geophys.
Res.
71,
3379
(1966);
J.
Chappell,
Geol.
Soc.
Am.
Bull.
85,
553
(1974);
and
H. H.
Veeh,
ibid.
89,
356
(1978).
13.
N.
J.
Shackleton
and
R.
K.
Matthews,
Nature
(London)
268,
618
(1977);
N.
J.
Shackleton,
Philos.
Trans.
R.
Soc.
London
Ser.
B
280,
169
(1977);
N.
J.
Shackleton
and
N.
D.
Opdyke,
in
(1).
14.
R.
G.
Fairbanks
and
R.
K.
Matthews,
Quat.
Res.
(N.
Y.)
10,
181
(1978).
15.
J.
T.
Andrews,
Prog.
Phys.
Geogr.
2,
375
(1978).
16.
L.
Heusser
and
N.
J.
Shackleton,
Science
204,
837
(1979).
17.
We
adopt
the
following
Pleistocene
classifica-
tion
of
D.
Q.
Bowen
[Quaternary
Geology
(Per-
gamon,
Oxford,
1978)]:
early
Pleistocene
1.61
to
0.70
million
years,
middle
Pleistocene
0.70
to
0.128
million
years,
and
late
Pleistocene
128,000
to
10,000
years.
18.
R.
Q.
Oaks,
Jr.,
and
J.
R.
DuBar,
Eds.,
Post
Miocene
Stratigraphy,
Central
and
Southern
Atlantic
Coastal
Plain
(Utah
State
Univ.
Press,
Logan,
1974),
p.
275.
19.
R.
Q.
Oaks
and
N.
K.
Coch
Science
140,
979
(1963);
J.
Hoyt,
Nature
(London)
215,
612
(1967);
D.
J.
Colquhoun
and
W.
Pierce,
Quater-
naria
20,
35
(1971).
20.
W.
F.
Ruddiman
and
A.
McIntyre
Science
204,
173
(1979).
21.
D.
J.
Colquhoun,
in
(18),
pp.
179-190;
J.
R.
Du-
Bar,
H.
S.
Johnson,
B.
Thom,
W.
0.
Hatchell,
in
ibid.,
pp.
139-173.
22.
T.
M.
Cronin,
Quat.
Res.
(N.
Y.)
13,
213
(1980).
23.
G.
B.
Shattuck,
Maryland
Geological
Survey,
Pliocene
and
Pleistocene
(Johns
Hopkins
Press,
Baltimore,
Md.,
1906),
pp.
21-137;
C.
W.
Cooke,
J.
Geol.
7,
577
(1930);
W.
T.
Ross,
P.
J.
Ross,
D.
J.
Colquhoun,
Palaeogeogr.
Palaeo-
climatol.
Palaeoecol.
9,
77
(1971).
24.
J.
H.
Hoyt
and
J.
R.
Hails,
Quaternaria
51,
(1971);
C.
D.
Winker
and
J.
D.
Howard,
Geol-
ogy
5,
123
(1977);
V.
A.
Zullo
and
W.
B.
Harris,
in
Structural
and
Stratigraphic
Framework
for
the
Coastal
Plain
of
North
Carolina,
G.
R.
Baum,
W.
B.
Harris,
V.
A.
Zullo,
Eds.
(Caroli-
na
Geologic
Society
and
Atlantic
Coastal
Plain
Geologic
Society,
Wrightsville
Beach,
N.C.,
1979),
p.
31.
25.
H.
G.
Richards,
Trans.
Gulf
Coast
Assoc.
Geol.
Soc.
19,
601
(1969).
26.
J.
E.
Hazel,
U.S.
Geol. Surv.
Prof.
Pap.
704
(1971);
U.S.
Geol.
Surv.
J.
Res.
5,
373
(1977);
T.
M.
Cronin
and
J.
E.
Hazel,
U.S.
Geol.
Surv.
Prof.
Pap.
112S-B
(1980).
27.
J.
L.
Liddicoat,
B.
W.
Blackwelder,
T.
M.
Cro-
nin,
L.
W.
Ward,
Geol.
Soc.
Am.
Abstr.
Pro-
gram
11,
187
(1979).
28.
J.
K.
Osmond,
J.
R.
Carpenter,
H.
L.
Windom,
J.
Geophys.
Res.
70,
1843
(1965);
W.
S.
Broeck-
er
and
D.
L.
Thurber,
Science
141,
58
(1965);
J.
K.
Osmond,
J.
P.
May,
W.
F.
Tanner,
J.
Geophys.
Res.
75,
469
(1970).
29.
B.
J.
Szabo
and
J.
N.
Rosholt,
J.
Geophys.
Res.
74,
3253
(1969).
30.
A.
McIntyre,
W.
F.
Ruddiman,
R.
Jantzen,
Deep-Sea
Res.
19,
61(1972).
31.
N.
G.
Kipp,
Geol.
Soc.
Am.
Mem.
145,
3
(1976).
32.
G.
M.
Peterson,
T.
Webb
III,
J.
E.
Kutzbach,
T.
van
der
Hammen,
T.
A.
Wijmstra,
F.
A.
Strut,
Quat.
Res.
(N.
Y.)
12,
47
(1979).
33.
L.
D.
Delorme,
S.
C.
Zoltai,
L.
L.
Kalas,
Can.
J.
Earth
Sci.
14,
2029
(1977).
34.
J.
E.
Hazel,
J.
Paleontol.
42,
1264
(1968).
35.
P.
C.
Valentine
[U.S.
Geol.
Surv.
Prof.
Pap.
683-D
(1971)]
describes
ostracodes
from
loca-
tion
3.
T.
M.
Cronin
[Geogr.
Phys.
Q.
33,
121
(1979)]
describes
ostracodes
from
localities
1,
2,
6,
8,
and
12.
Species
lists
are
available
for
other
localities
from
T.M.C.
36.
T.
M.
Cronin,
Quat.
Res.
(N.
Y.)
7,
238
(1977).
37.
0.
Elofson,
Zool.
Bidr.
Uppsala
19,
215
(1941);
T.
M.
Cronin,
unpublished
ostracode
occur-
rence
data
from
Continental
Shelf
off
South
Car-
olina,
Georgia,
and
Florida.
38.
J.
E.
Hazel,
U.S.
Geol.
Surv.
Prof.
Pap.
529-E
(1970).
39.
Assemblage
from
locality
6
contains
the
sub-
tropical
species
Neocaudites
atlantica
Cronin,
1979,
Proteoconcha
tuberculata
(Pun,
1960),
Paracytheridea
altila
Edwards,
1944,
Pellucis-
toma
magniventra
Edwards,
1944,
Hulingsina
glabra
(Hall,
1965),
Puriana
convoluta
Teeter,
1975,
and
others,
but
also
contains
rare
speci-
mens
of
Finmarchinella
finmarchica
(Sars,
1865),
Leptocythere
angusta
Blake,
1929,
Muellerina
canadensis
(Brady,
1870),
Lox-
concha
sperata
Williams,
1966,
and
Ben-
sonocythere
americana
Hazel,
1967.
All
are
temperate
species
whose
presence
may
signal
slight
cooling
at
this
locality.
40.
D.
G.
Frey,
Ecology
32,
518
(1951).
41.
__,
Am.
J.
Sci.
25,
212
(1952).
42.
D.
R.
Whitehead
and
J.
T.
Davis,
Southeast.
Geol.
10,
149
(1969).
43.
G.
T.
Trewartha,
A.
H.
Robinson,
E.
H.
Ham-
mond,
Elements
of
Geography
(McGraw-Hill,
New
York,
1967).
44.
D.
R.
Whitehead,
Ecology
44,
403
(1963).
45.
__,
in
Quaternary
Paleoecology,
E.
Cushing
and
H.
Wright,
Eds.
(Yale
Univ.
Press,
New
Haven,
Conn.,
1967),
p.
237.
46.
__
,
Quat.
Res.
(N.
Y.)
3,
621
(1973).
47.
A.
E.
Radford,
H.
E.
Ahles,
C. R.
Bell,
Manual
of
the
Vascular
Flora
of
the
Carolinas
(Univ.
of
North
Carolina
Press,
Chapel
Hill,
1968).
48.
R.
G.
Johnson
and
B.
T.
McClure,
Quat.
Res.
(N.
Y.)
6,
325
(1976);
J.
Andrews
and
R.
G.
Bar-
ry,Annu.
Rev.
Earth
Planet.
Sci.
6,
205
(1978).
49.
W.
F.
Ruddiman,
personal
communication.
50.
P.
J.
Fox,
B.
C.
Heezen,
G.
L.
Johnson,
Science
170,
1402
(1970);
P.
A.
Rona,
Geol.
Soc.
Am.
Bull.
84,
2851
(1973);
W.
C.
Pitman
III,
ibid.
89,
1389
(1978).
51.
J.
F.
Marshall
and
B.
Thom,
Nature
(London)
263,
120
(1976).
52.
N.
J.
Shackleton,
in After
the
Australo-
pithecines,
K.
W.
Butzer
and
G.
L. L.
Isaacs,
Eds.
(Mouton,
The
Hague,
1975),
p.
1.
53.
R.
K.
Matthews,
Quat.
Res.
(N.
Y.)
3,
147
(1973).
54.
M.
L.
Bender,
R.
G.
Fairbanks,
F.
W.
Taylor,
R.
K.
Matthews,
J.
G.
Goddard,
W.
S.
Broeck-
er,
Geol.
Soc.
Am.
Bull.
90,
577
(1979).
55.
A.
L.
Bloom,
ibid.
78,
1477
(1967).
56.
R.
I.
Walcott,
Quat.
Res.
(N.
Y.)
2,
1
(1972).
57.
J.
A.
Clark,
W.
E.
Farrell,
W.
R.
Peltier,
ibid.
9,
265
(1978).
58.
N.-A.
Morner,
Ed.,
Earth
Rheology,
Isostasy
and
Eustasy
(Wiley,
New
York,
1980),
p.
599.
59.
B.
W.
Blackwelder,
0.
H.
Pilkey,
J.
D.
Howard,
Science
204,
618
(1979);
D.
W.
Bowen,
Nature
(London)
283,
619
(1980);
J.
T.
Hollin,
ibid.,
p.
629.
60.
J.
K.
Osmond,
J.
P.
May,
W.
F.
Tanner,
J.
Geophys.
Res.
75,
469
(1970).
61.
T.-L.
Ku,
K.
G.
Knauss,
G.
G.
Mathieu,
Deep-
Sea
Res.
24,
1005
(1977).
62.
Oceanographic
Atlas
of
the
North
Atlantic
Ocean,
section
II,
Physical
Properties
(U.S.
Naval
Oceanographic
Office
publication
700,
Washington,
D.C.,
1967),
p.
300.
63.
We
thank
H.
Markewich,
K.
R.
Lajoie,
A.
L.
Bloom,
W.
F.
Ruddiman,
and
J.
T.
Andrews
for
reviewing
a
draft
of
the
article,
R.
Mixon
for
helpful
comments,
E.
Compton
for
drafting
the
figures,
and
L.
Miller
for
typing.
SCIENCE,
VOL.
211
240