Geochimiea et Cosmoehimica Acta 1962, Vol. 28, pp. 1757 to 1786.
Carbonand oxygen isotopic composition of mollusk shells from marine and fresh-water environments*
M. L. KEITH, G. M. ANDERSON and R. EICELER~
Dept. of Geochemistry and Mineralogy, The Pennsylvania State University, Universit-yPark, Pa.
(Received December 1963)
Abstract-Results are given of a system&c survey of differencesin the isotopic composition of carbon and oxygen of modern mollusk shells from m&ne and continental enviromnents. Marine shells analyzed show a range of 6Cr3 (relative to Chicago PDB standard) from +42 to - 1.7 %, whereas the fresh-water mollusk shells have relatively Cls-descient carbon, in the rage BY = -0-6 to - 152. There is a similsr differencein Or* content.
Within the marine group, environmental sub-groups differ mainly in 01s content and the differencesare consistent with the temperature dependencepreviouslystudied by other investigators. Within the fresh-water group, the most striking differencebetween sub-groups is in the Cl3 content of pelecypod shellsfrom large lakes (6Cr3 = -2.4 to 6.07&J,and from rivers (NY = -8.3 to - 15.2 X), a differencewhich is shown to be environment-controlledratherthan speciescontrolled. The soft parts of pelecypods also show characteristicdifferences of carbon isotopic composition from marine to lacustrine to fluvial specimens. The ligament of peleoypods is found to consist partly of 8ragonite fibers which are isotopically different from shell carbonate.
It is concluded that the carbon isotope ratio in mollusk shells is considerablyinfluencedby the proportional amount of land-plant derived carbon included in the food of the mollusks or contributed by humus decay to dissolved bicarbonate in the water. It appears likely that the observed isotopic differences can be applied to the env~o~ent~ study of fossils and sedimentary carbonate rocks, 8nd to the source identiEcation of shell artifacts of archaeological interest.
THE environment of deposition of sedimentary rocks and the extent of the oceans and continents during different geologic periods has been established by the study of fossils and of regional changes in lithology. In sections where fossils are scarce, or where continental and marine beds are intercalated in a cyclic or irregular way, the problem of determining depositional environments becomes complex and it is desirable to have criteria which will supplement those based on the fossil record,
Investigations of the use of isotopic criteria for differentiating marine and freshwater carbonate rocks have been carried out at this laboratory for several years, beginning with the work of CLAYTONand DEGENS (1959). The present study of modern mollusk shells from known marine and continental sites was undertaken as a basis for understanding and interpreting variations in the carbon and oxygen isotopic composition of fossils and limestones in relation to their environment of deposition.
STI~UCTURE AND FORMATION OF MOLLUSK SHELL
Aquatic mollusks occupy a wide variety of environments, from marine to brackish to fresh-water, and tropical to arctic, and they exhibit a wide variety of preferences
* Contribution 52-117 from the College of Mineral Industries, The Pennsylvania State University.
t Present address: Geologisch Institute, Rhein. Friedrich-W&elms Universitat, Bonn, Germany.
1557

1758

?rl. L. KEITH, G. M. ANDERSONand R. EICHLER

for different bottom conditions; many live in clean sand, others are mud dwellers, still others prefer to attach themselves to rocks exposed at low tide.
Most bivalves feed on minute plants, diatoms and protozoa ; some, such as oysters and Venus clams are suspension feeders, while others, such as Tellina and Inacorna, are deposit feeders. Among the gastropods a considerable proportion are carnivorous, some are deposit feeders, others live on aquatic plants (ABBOTT, 1953). The shell of the mollusks is deposited by the mantle in successive layers with a micro-architecture which is characteristic of the genus (BOGGILD, 1930). Most marine shells and all of the fresh-water shells examined consist mainly of aragonite. Some marine species
deposit a calcite shell, and a few, such as Jfytilua and Chama, deposit an outer layer of calcite and an inner pearly layer of aragonite.
Pelecypods probably deposit shell mainly during the warmer part of the year, above some minimum water temperature which varies with species. Many gastropods, on the other hand, show evidence of shell deposition over a wide range of water temperatures (EPSTEIN and LOWENSTAM, 1953).

OXYGEN AND CARBON AVAILABLE TO MOLLUSKS
(a) Oqgen
Variation of the 018: 016 ratio in the hydrosphere depends mainly on the fact that the vapor pressure of H,016 is greater than that of H,Ol*. Oxygen-18 is relatively concentrated in ocean water and varies with salinity over a range of about 0+3”/0(EPSTEIN and MAYEDA, 1953). In comparisonwith the ocean, continental waters. are relatively deficient in oxygen-18 and isotopically more variable. EPSTEIN and MAYEDA found a depletion of oxygen-18 in continental waters at progressively higher altitude and higher latitude, over a total range of about 5% in 0’8 content. The temperature dependence of oxygen isotope fractionation in exchange reactions between calcium carbonate and water, based on calculations of UREY (1947) was determined experimentally for inorganic carbonate precipitation (MCCREA, 1950) and subsequently for organic precipitation by marine organisms (UREY et al., 1951). The amount of isotope fractionation changes with water temperature so that for example, at 25°C the 018/01a ratio is higher by 2.85% in the calcium carbonate and at 7°C by 3.27% than it is in the water (CLAYTON, 1961).

(b) Carbon
The isotopic composition of carbon in the various sources available to mollusks, is pertinent to the present study. Omitting marine limestones, with KY3 near zero (Chicago PDB scale), the principal inorganic reservoirs or sources of carbon, in order of decreasing carbon-13 content are: (1) ocean-water bicarbonate, with 6C13 about -2 per mil, (2) atmospheric carbon dioxide with KY3 about -7 per mil, and (3) fresh-water bicarbonate with widely variable 6C13, generally less than -8 per mil.
REVEL~E and FAIRBRIDGE (1957) quote HARMON CRAIG to the effect that marine invertebrate tests appear to contain a mixture of carbon derived from metabolic activity and carbon derived from sea water bicarbonate. WILBUR (1960) concludes that the metabolic pathway is the more important one in mollusks.

Carbonand oxygen isotopiccompositionof molluskshells

1759

PREvIOnS ISOTOPICMEASUREMENTS ON MOLLUSK SHELLS
Oxygenisotopic data for mollusk shells (fossil or modern) have been reported by UREY et aE.(1951), EPSTEINet uZ.(1953), EPSTEINand LOWENSTA~(~1953), CLAYTON and DE~ENS(1959) and KEITH,EIC~ER and PARKER(1960). The first reference and the last two, above, include data on carbon as well as oxygen.
Investigations of the carbon isotope ratio alone, without reference to oxygen have been reported by CRAW (1953, 1954), JEFFERYet al. (1955) and BROECKERand OLSOK (1961) for marine mollusk shells and fossils. Limited data for fresh-water shells are given by BROECKERand WALTON (1959), BROECKERand OLSON(1959), OANAand DEEVEY(1960) and KEITE and ANDERSON(1963).
The data, summarized by GEAF(1960), generally confirm the suggestion of UREY (EM?) regarding oxygen, and the observation of CLAYTONand DEQEXS(1959) to the effect that carbon and oxygen isotope ratios can be used with some caution, as supplementary criteria to differentiate marine and fresh-water carbonates. Understanding and use of the isotopic criteria for interpreta$on of sedimentary environments requires that a background be established by two additional stages of investigation: (1) detailed analysis and comparison of modern carbonate samples collected from known environments, and (2) statistical studies of fossils and limestones of different geologic ages. The present paper is a contribution to the first stage.

SAMYLETREATMENTAND ANALYTICALPROCEDURE

With a few exceptions, ‘notably the suite from Cape May, New Jersey, mollusk

specimens were collected alive. Most of the shell samples were simply air dried after re-

moval of soft parts. A few specimens were preserved in diluted isopropyl alcohol or in

formalin for a few months. Remnants of soft parts were removed by scraping and by

treatment in commercial Clorox ( -5% solution of sodium hypoehlorite). Formalin

has a slight etching effect on shells; however, it was shown that preservation or

treatment with isopropyl alcohol, formalin, or Clorox have no measurable effect on

the carbon or oxygen isotopic composition of shells (EPSTEINand LOWENSTAM1,953;

EICHLER,1961).

In general, large shell specimens were sampled by sawing off a diagonal slice to

represent a number of growth layers. The heavy hinge region of pelecypod shells was

avoided, except for very small specimens, which were sampled in toto. Samples for

investigation of within-shell variation were obtained by sawing or grinding off the

desired portions.

Cleaned carbonate material was crushed to -80 mesh and heated for 20 min at

42O*Cin flowing helium, to remove or pyrolize organic compounds. The resulting

carbonate residue was treated with 100% phosphoric acid in evacuated tubes, and

carbon dioxide evolved over a 24 hour period. The carbon dioxide was purified

and collected and was analysed with a 6 in., go-degree sector mass spectrometer,

following the procedure of MCKINNEYet al.(1950), in which the isotopic ratio of the

sample gas is compared with that of a carbon dioxide standard. Carbon isotopic

compositions are expressed as SCla, the difference, in parts per thousand, between

the carbon-13 oontent of the sample gas and the Chicago PDB standard carbon

dioxide.

KY3 = 1000

(yS'(-y

- (-y/p

I ssmpie

SM.

c=/c;;fd.

1760

M. L. KEITH, G. M. ANDERSON and R. EICHLER

Oxygen isotopic compositions are expressed similarly, as SO18,in terms of the 018/0f6 ratios 0f 9amples and standard. Data are corrected for the effect of 0” on the carbon isotope measurement and for the effect of Cl3 on the oxygen isotope measurement (CRAIG, 1957).
Duplicate carbon dioxide sub-samples were prepared and analysed at different
times with a total analytical random error of less than 0.2 per mil. Reproducibility of the measurement on a single carbon dioxide sample is somewhat better than -&Owpl er mil. Recorded &values are the means of replicate runs.

ALLEGHENY R., GRAND R., MICH: MERAMEC R., MO

T‘ENNESSEE R., ALA.
-8 -
GREAT LAKES--y&
-4-

.4 CALIFORN +2

MARINE --- + 9

60’8, x.

-4

-8

-12

Fig. 1. Isotopic composition of marine and fresh-water pelecypod shells. Symbols within open circles indicate samples from river sites. Open circles without. interior symbols indicate lacustrine samples, four of which on Fig. 1 and seven on Fig. 2 are from smali lakes on the Canadian shield. Half-filled circles (Big. 1) repre-
sent samples from intermediate salinity waters (Table 7). A = sub-Arctic samples
61-151, 152.

COMPARISOW OF MARINE AND FRESH-WATERSHELLS
Carbon and oxygen isotopic data for marine and fresh-water shells are given in Table 7 and shown graphically in Figs. 1 and 2. Information regarding the collection sites is given in Table 8.

For the readers who may wish to compare the data with results reported relative to other standards, the oxygen data can be recalculated to “Mean Ocean Water” standard by the formula of CLAYTON and EPSTEIN (1958);

60w = 1.0295(6,,,~7 29.8

Similarly, for recalcuiating Taothe ?U’BS standard (Solenhofen limestone):

Carbon BXyBss=o t-OOf &,a + 1.06

Oxygen 6X,szo = I+04 Gpns - 4.16

Carbon and oxygen isotopic composition of mollusk shells

-14*rdo-
8-6-

MARSHY

FS eF
OF

SMALL LAKES OF CANADIAN SHIELD

_ _--;;.

/

/

/c 5 0

CD

/

I

C’

‘e___zl_---’

1761

ARCTIC\
,% \ I ’ ’ c ‘yt

GASTROFWS [FRESH-WATER-- 0)

LA’JCLU

&/’

+

,

/

2

-4

-

-

_to

1
-e

bO*, %

Fig. 2. Isotopic composition of marine and fresh-water gastropod shells.

Symbols within open circles indicate samples from river sites. Open circles

without interior symbols indicate lacustrme samples, four of which on

Fig. 1 and seven on Fig. 2 are from small lakes on the Canadian shield.

A = sub-Arctic samples 61-151, 152; B =i Bermuda; F = Florida; SL =

Scammon Lagoon, Mexico.

Table 1. Carbon isotopic composition of mollusk soft parts in comparison with shell

Sample number

Species

Marine samples 62-132 62-133

Ti&u ~~~~~
&f$&i$ ~~f~~~~~

Laoustrine samples

Mean

Soft parts oC’3 (%a)
-17*0 ---16.4 -167

Shell SCr3 (%A
$1.12 $0.24 + 0.68

Difference (%A
18.1 16.6 17.4

61-12
61-22 61-24

Amblema CQ&&Z~&C& Elliptic &i&&w Elliptio wmplanatua

-22.6 - 24.7
- 25.4

- 2.44 -4.87
-4.52

20.2 19.8
20.9

Fluvial samples

M0FUl

-24.2

-3.94

20.3

61-202 61-239 61-240

Actinonaiaa oarinuta Strophitua rUgQsuS Strophitw TUgQ8U.9

-30.2 - 30.6 - 30.4

-11.50 -11.12
-11.74

18.7 195
18.7

61-241

Strophitw ~osua

-27.2

-11.44

15.8

61-242

Strophitw rzdgosw

-27.6

-11.76

15.8

Mean

-29.2

-11.51

17.7

Combustion and measurement of soft parts by J. r\‘.WEBER. Pine wood, measured at the same time, for comparison. gave &Cl3 = - 24.37&. hiBS 21 (graphite) gave &Cl9 = -26,57_ (cf. -278:&, measured by CRAIG, 1957).

1762

M. L. KEITH, G. M. ANDERSON and R. EICHLXR

Table 2. Isotopic composition of interior and exterior layers of marine pelecypod sheik

“Species”

Locality

Interior

6OL8“, /, EXV&lGX DIFF.

82-52 ostreo
82-132 Tivela

Banderas, Mex. Balboa, Cal.

In - 3.38 (-on?)*

EX - 3.96 (-1.19)

In - Ex + 0.59 f0.53

62-17 02-42 62- 55 m-73 62-62 62-133 68-262 62-211

charno Spondylua
Atrina Attine
&iytiaw
Mytilus M&Z?& aaytizus

Mazatlau, Mex. Cleopha, Mex. 0. of Mexico Breton Id., La.
La Joils. Cal,
La Jolla, Cd. Cape May, N. J. Hudson Bay, Can.

(- 1.85) (-2.11) (-1.33) (-2.24) (- 1.63) (-1.40) (- 1.46) (- 3.97)

-2.39 -2.72 -1.21 - 1.37 - om -0.90 -O*BO -3.17

3-0.54
j-061 -0.12 -0.87 -0.74 -0.50
-0+.6 -0.80

Mean Difference (without regard to sign)

0.65

* Parentheses indicate aample composed of aragonite; others are calcite.

SsciS%, *

Interior Exterior DIFF.

In -0~20
(+ 1.54)

+%3 (+0.66)

In - Es - 0.83 + 0.8’3

(+ 1.28) ($1.02) (f 1.72) (+oW (+@16) (+6-34) if 1.13) (-0.07)

+0.38 +0.71 + 0.99 + 0.46 - 0.38 to.14 - 0.58 - 0.84

+-0.90
-+ 0.31 +0.73 + 0.37 + 0.54 +0*20 + 1.71 +a*77

0.73

Table 3. Isotopic composition

of marine peleoypods: within-shell with ligament fibers

variation

and comparison

Sample number

Species

Sub-eample distance from

Site

beak

601’

6C’”

(mm)

(%o)

%I)

62-62

My&h% ca1q0WGWA.9

La Jolla, Calii. Exterior (calcite):

30

-0.47

-0.31

48

-1-12

-0-28

67

- 0.83

-0.06

86

-0.94

-0-09

99

- 0.70

-O*Bi’

115

- 1.28

-0.58

62-132

Interior (aragonite) Ligament fibers (aragonite)
Balboa, Calif. Exterior (aragonite)

Mean
18 36 54 73 93 112

-0.89 -1.63 - 1.59
-1.51 -0.84 -160 -1.13 -1.06 -1.60

-0.38 +0*16 + I.89
+ 0.54 + 0.68 +067 +0.68 +o-75 10-61

62-52

Interior (sragonite)

Ligament fibers (aragonite)
-

Ostrea in’descena

Banderae Bay, Jllexico

Shell (calcite and aragonite)

Ligament fibers (aragonite)

62-133

1Mytilu.3 u&lifontianua

La Jolla, Cal.

Shell (calcite and aragonite)

Ligament fibers (aragonite)

Average difference: (ligament fiber &mean shell 6)

Mean

-1.19 -0.76 -0.28
-3.67 -2.54
- I.15 -0.76
o-47

+0.66 + I.54 + 2.42
10.21 +0.25
+ 0.24 +1.94
1.21

Carbon and oxygen isotopic composition of mollusk shells

1763

Table 4. Isotopic composition of fluvial peleoypods: within-shell variation and comparison with ligament fibers

Sample number

Species

Site

Sub-sample distance from
beak
(mm)

&Y* (%J

&?a (%0)

61-101

Negalonaias gigantea Tennessee R. Alabama

Exterior:

37

-7.39

-11.48

50

--5.12

- 9.30

58

- 7.12

- 9.88

66

-6.71

- 9.71

83

- i-16

- 9.73

91

- 7.33

- 994

97

-7.20

- 9.85

104

-6.88

- IO.20

-7.11

- lO*Ol

Interior: Muscle f3car Middle Lip
Ligament fibers

Mean

-6,56 -7.08 -7.49
-7.04 -6.42

-648 -9.73 -9.92
-9.38 -8.10

61-127

Actiwias cminata

French Creek, Pa. Exterior:

Interior: Lip Hinge

20 33 49 60 i3 88
Mean

-9.44 -3.22 -890 -9.60 - 960 -9.33
-9.35

- lo,14 - lo*01
Mean - 10.08

-11~92 -11.34 -1169 - 12.79 -14.61 -16.56
-13.15
-16.25 - 15.62
- 15.93

Ligament fibere

61-2

Ad&mu

oos6ccda

Grand R., Mich.

Shell

Ligament Fibers

6248

Ligumka recta

Meramec R., Missouri Shell Ligament fibers

Average difference: (ligament fiber &mean shell 6)

- 9.02

--x3*79

- 8.55 -8-12

- 12.45 -11+65

-7.12 -6.72
0.54

-1391 -13.10
1.01

1764

M. L. KEITH, G. M. ANDERSON and R. EICHLER

Table 5. Variation of shell isotopic composition within communities of fresh-water pelecypods

(a) French Creek, north of Meadville, Pa.

Species Actinonaiaa car&&z

Sample number
61-192 61-193 61-194 61-195 61-126 61-196 61-197 61-198 61-199 61-200 61-202 61-201 61-127

Weight of single valve
(g)
1.4 2.4 4.9 7.7 15.0 17.9 21.0 36.7 42.8 68.2 82.4 91.4 122.0

6018
(%a)
- 10.34 -10.24 - 10.07 -10.13 -10.10 - 10.03
-9.91 - 10.05 -10.01 - 10.07 - 10.07
-9.76 - 10.14

603
(X)
- 10.96
-11.29 - 10.53 -11-56 -11.18 -11.66 - 12.33 - 13.77 - 12.26 -14.72 - 12.26 - 14.03 - 16.25

Elliptio dilatatus (purple shell)

Mean
61-177 61-178 61-181 61-182 61-183

6.5 7.8 13.9 19.2 19.8

- 10.07
-10.31 -10.11 - 10.20 -10.18
-9.95

-12.52
- 12.41 -11,99 -11.97 - 14.04 - 14.96

Larnpsilia faaciola

Mean
61-184 61-188 61-185 61-186 61-187

13.4 14.1 14.6 16.0 22.4

- 10.15
-9.97 -9.73 -9.93 -9.81 -9.70

- 13.07
- 13.39 - 12.00 - 12.37 - 14.02 - 13.57

Pleurobenau coro!atum coccineum (pink shell)

Mean
61-189 61-190

4.4 14.1

- 9.83
- 10.26 - 10.29

- 13.07
- 10.43 -11.57

Pleurobema cordatum coccineum (white shell)

Mean
61-124 61-191

16.0 30.6

- 10.27
- 10.25 - 10.20

-11.00
- 10.67 -11.51

Lasnaigona costata

Mean

61,123B

25

61-123A

50

- 10.22
- 9.94 -9.59

-11-09
- 13.07 - 13.18

Lamps%8 ovata venttio.9a

Mean

61-125

65

61-119

118

Mean

French Creek, mean of six species

-9.77 -9.54 -9.36
-9.45 -9.92

-13.12 -11.03 -12.03
- 11.53 - 12.39

Carbon and oxygen isotopic composition of mollusk shells

Table 5. cc&. (bf &and River at Dunnw’lZe,Q&c&o
Species

Sample number

Quadrula quadrula

61-211 61-212 61-209 61-210
Mean

Weight (8)
29.7 32.4 46.3 47.3

6018
(%a) -
- IO.62 - 10.57 -10.63 - 10.42
- 10.56

1765
I333 (X0) -10.97 -11.85 -14.30 -13.63 - 12.69

Cumberlund~ rnmodonkz

61-232 61-231 61-229 61-228 61-230 61-227 61-226 61-224 61-223 61-226 61-222 61-221
Mean

1.5 5.9 7.2 7.4
7.5 9.9 13.0 13.2 19.8 21.5 32.8 43-6

-7.9%
-7.99
-7.69 -7.68 -7.50 - 7.51 -7.66 -7.77 -7.52 -7.35 -7.56 - 7.43
-7.64

- 12.27 -11.44 - 11.48 -11.46 - 10.80 -11.28 -12.03 -11.64 -11.34 -11.31 -11.15 - 10.50
-11.39

The data above are for samples of recently deposited calcium carbonate, from the lip of each shell.
Table 6. Isotopic composition of aragonite fibers from the ligament of fluvial pelecypods*

Sample number

Weight of single valve
(g)

61-192 61-193 61-195 61-196 61-197 61-198 61-199 61-200 61-202 61-201 61-127

1.4 2-4 7.7 17.9 21.0 36.7 42.8 68.2 82.4 91.4 122.0

Mean
Mean difference from lip shell of the same specimens (Table 5), ligament S-shell S =

6018 (“/,)
-9.38 -9.55 - 10-33 - 12.26 - 9.72 -9.46 -9.35 - 9.08 -9.18 -9.11 -9.02
- 9.68
10.3%

NY (?A)
-11.66 -11.04 - 12.15 -11.37 -11-02 -12,IO -11.57 - 12.20 -11.50 - 12-70 - 13.79
-11-92
iO.78

* Size sequence of Actinmiaa cati~

collected alive from a single

community: French Creek, north of Meadville, Pa. (see Table 5 and Fig. 6).

1766

M. L. KEITH, G. 31. ANDEWOS

and H. EICHIXK

Location

Sample number

Table 7. Isotopic composition Species

of mollusk shells
taco, struct.

00 ’ 8
(“A”)

,)C“” ( “A)

(a) M.ARINE SAMPLE

(1)

56-641

58-257

58-258

58-259

58-260

.58-261

X3-262

Cape May, ,‘?ew Jersey (.Mnntic)
Spisula solidis.~ima Dillwyn Tqelus plebe&s Solander Ensis directus Conrad Mactra sp. _Voetia. ponderosa Say &quip&en irrcrdians Lamarck MytilU8 e&&s L.

Ca.pe May, mean

Shallow Nesican Waters (Pucific)

(2) _62-2t
62-4 62-6

Turritella gonosloma Valenciennes Fissurella vim&ens Sowerby Crucibulum scutellatum Wood

62-l 1

Cerithium mnculosum Kiener

62-13

Trachycard,iUm senticosum Sowerby

62-14

Anadara esmeralda Pilsbry and Olsson

62-16

Pseudochama cowugnta Broderip

62-17

(3)

_62-28

62-32

Chanra sp. (piece of large shell) Strombua galentua Swainson Cerithium maculosum Kiener

62-34

Codakia distinguenda Tryon

62-38 62-40

Pseudochama corrugata Broderip Fi.wurella viriscens Sowerby

62-42

Spondylus princeps Broderip

(4)

62-52

Ostrea iridescens Gray

(5)

62-122

Crucibulum spinosum Sowerby

a

62-123

Crepidula atriolata Menke

a

Shallow M&can

Waters, mean Pacific Ocean, 30 to 90 meter8

(6) 62-18

Cantharus cf. capitaneua Berry

a

(7)

62-19

Cantharus cf. cupituneus Berry

a

62-20

Chione kelletti Hinds

a

(8)

62-22

(9)

62-23

62-24

Crucibulum sp. nav.

a

Murex (Hexaplex) brassicn Lamarck

a

Fusinus panamensis Dall

a

(10)

62-43

Buraa nana Sowerby

a

62-44

Chlamys circularis Soaerby

c

(11)

62-190

Lunatia lewisi Gould

a

62-192

Kellelia kelletz Forbes

a

(12)

62-193

62-194

Xyti1u.s californianus Conrad Kelletia kelleti Forbes

6--2-195

Burma californica Hinds

Pa&c 0 aean, 30 to 90 meters, mean

Deep Sea (Pnci’c, 3000 meters)

(13)

62-25

62-26

Dentalium megnthyris Dal1 (Scaphopod) Limo@8 comprensus

Deep Sea, mean

* a = araxonite,

c = calcite.

t Gastropoda indicated by sample number underline.

Det,ails regarding

-1.53 -2.2s ._ 0.93 -?.3li --_1.68 _ “.‘,j ~-0_ .91, - 1.83
-2.43 -1.90 - 1.99 --I.31 -_“..iz
- 1.94 - 2.84 -2.1’7
-1.96 - 1.9s -2.4; -2.3” _ 1.91, -_7.*2 - 3.00 -0.46 - 1.50 -2.1”
-0-80 -0.16 -0.6” - 0.02 - 1.34 - 0.08 -1.13 -0.63 i-o.44 10.24 -O.l? $0.63 + 0.08 -0.2;
i2.96 T 2.4; _“.;.’

-0.91 -- 0.47 - 11.23 - (I.33 - 1.03 - 10.64 _‘0__.10 - 0.06
-.. Z.‘i._’ ..~2.03 _~ 1.26 -.~I.61 _ 1.25 -_+.j5 __ 1.jg -- 0.83 -.- 1.34 - l.i4 - 1.41 __ l..jz -_4.“3 -- O.Y7 -0.04 -- 0.89 -- 0.90
- 1.33
---0.90 -0.74 -0.14 ---I).89 -0.29 -- 0.96 -0.52 - 0.05 - 0.30 __ 1.36 .:. 0.20 --W-56 -~ 1.64 __~().47
-1.13 -- 0.?4
--O.ti!)

zites arc given in Table 8.

Carbon and oxygen isotopic composition of mollusk shells

1767

Table 7 con&

Locat.ion

Sample number

Species

taco, struct.

6018 ix,

&C’S (X)

(14)
(15)
(17) (18) (19) (20) (21)

62-49
62-61, 62, 133 62-64
62-74 62-184 62-185 62-196
62-132

La Jolla and Newpvrt, California (Pacify)
Donax gouldi Dab
Mytitw californianis Conrad (mean of 3) 01iv&a biplicata Sowerby Maccnna eecta Conrad Leptopecten latiauraruaConrad Nassatiaa?fO86&W Gould
H&o& fdgensPhilippi (bulk ssmple)*
Tivekz ~~~~~~~ Mswe (bulk ssmple)*

a
a+c
B
&
c a a a

La Jolla and Newport, mesn

Canadtin Arctic

62-230
62-231 62-232 62-233 62-214
_62-217 62-215

Macwna c&area Qmelin Clinacardiuntciliaturn Fabricius Hiutella amtica L.
Portlandia arcticu Colus ap. Buccinunz sp. Nuculana sp.

Canadian Arctic, mean

56 Marine semples, menu

-1.39 -1.15 -0.87 -0.69 -0.68 -0.69 -0.60 -0-76
-0-91
+ 1.36 +0.45 10.94 +0.41 +0.86 +0*92 -to.35 $-o-75
-0-91

+I.86 +0*33 j-O*67 -t-1*55 +1*05 +0*10 -f- 0.63 t-1-12 -to40
-9.37 -0.15 11.26 -1.67 -0.36 +1*07 -0.69
-0.14 = + 0.68

(b) SPECIMEKW FRO=

(161
(22) (23)

62-211 62-197 62-212 62-213

33%i%~mwm

-!rY

wnTgast

&f$<l_ e&&e Linn6
08trea sp.
Mytilu.~ edu& L. Macoma b&him L.

e+a
c
C+B
a

- 357 -4.12 -6.64 -6.11

-0.46 -183 - I.94 -2.60

(c) FRESH-WATERSAWPLEG

Sdl lakur, ponda and baya

(25) (26) (27) (28) (29) (30) (31) (32) (34) (35)

56-644 61-40 61-42 61-46 61-48 61-49 61-50 61-51
61-53 61-138 61-138A 61-139

EUiptio com.planatw Dillwyu Aacatmtorbiebahien& Dunker Ampullaria pa&&m Say Physa parkeri “Currier” de Camp Physa ~emstmp~ Say Lynwzaeu ernarginataangt&ta Sowerby He&nxa trivohia t&x&i8 Say Pl~norb& a97nigera Say Hydrvbia bermuduensis Pilsbry Stagnicola pa&t&s elodee Say
Aplexu hypnomrn L. Heliavma trivolvis trivolvis Say

-9.57 -3.45 -4.86 -7.04 -11.81 -6.75 -5.13 - 8.67 - 3.95 -8.80 -9.39 -6.75

-9.82 -5.68 -9.14 -2.10 - 12.07 -3.75 - 4-02 - 9.40 - 10.81 - 10.79 -11.93 -9.28

* Uncorrected radiocarbon ages: Five& 40 -& 150 years; Hal&is, 125 f 150 years (Michigan analyses M-1221 and 1222), KEITH and ANDERSON(1963).
t Selinit.y ranges at collection sites: Lee. (16) Hudson Bay 22%, to 27$$, Lot. (22) Hood Canal 5:& to 300/, Lot. (23) James Bay loo/, to 207&,
2 -411of t.he fresh-water moIlusk shells are composed of arsgonite, therefore crystal structure is not indicated for individual samples.

1768

&f. L. KEITH, G. M. ANDERSON and R. &XLER

Table 7 cont.

Location

Sample number

Species

80’8

c)C’3

Small lakes, ponds and bays cont.

(36)

61-140A

61-140B

61-141

Helisoma trivolvis trivolvis Say Helisoma anceps Menke Valvata tn’can’nata Say

61-142

Anodonta grandis Say

(37) 61-143

(38)

61-149

61-150

Anodontu grandis Say Stag&cola cf. palusttis elodes Say Stagnicola cf. emarginata ontarienais Kiister

(39) 61-151

(40)

61-152

(41)

61-154

(42) 62-201

Sphoerium nitidum Clessin Discus or Zonitoides sp. Goniobasis livescens Menke Elliptio complanatus Dillwyn

Small lakes, ponds and bays, mean

Great Lake8 Suite

(43)

56-639,640

(44) 61-12

61-14

Qoniobasis livescens Menke, mean of 2 Amblema costata plicata Say Proptera alata Say

61-15

Lampsilis ventricosa Barnes

61-16

Lampsilis siliquoidea Barnes, (bulk sample)*

61-17

Pleurobema cord&urn coccineum Conrad

61-18

Elliptio dilntatus Ref.

61-20

Fusconnia &VU Ref.

(45) 61-21

Lampsilis siliquoidea Barnes

(46) 61-22 61-23

Elliptio dilatatus Raf. Lampsilis radiata Gmelin

61-24

Elliptio complanatus Dillwyn

61-25

Anodonta grandis Say

(47) 61-85

(48)

61-l i6

(49) 62-200

(50) 62-202, 203

Larnpsilis radinta Gmelin Pleurocera acuta Raf. Proptera alata Say Lampsilis siliquoidea roaacea, mean of 2

Great Lakes suite, -Mean

Rivers and Stream8

(51)

61-I

61-2

61-3

61-4

61-3

61-6

(32)

61-i

61-8

61-9

61-10

61-11

Elliptio dilatus Raf. Amblema cost&a Raf. Cyclonnias tuberculata Ra.f. Fwconaia j&a Raf. Quad&a pustulosa Lea Lampsilis ventricosa Barnes Fusconaia $ava Raf. Stroplitua rugosus Swainson Lnsmigona costata Raf. Quadrulo puatulosa Lea El!iptio rlihtrttus Raf. (bulk sample)t

- 10.54
- 10.43 - 10.66 -11.15 -11.79 - 10.40
-9.84 -1.5.21 - 17.50 - 10.03
-9.86
-9.29

-11.04 -11.14 -11.08 - 12.04
- 9.48 - 6.05 - 9.35 -4.11 -3.91 __ i.42 -11.33
--8.51

- i.42
-7.66 _ 7.94
-7.il - i-59 - 7.74 - _1.-1_J - 7.90 _ 7.20 -8.52
- 8.49 - 7.20
- 6.56 _ 7.80 - 7.2;
- 7.20 - 7.64
-7.62

-- 0.64 -2.44 --J-’, c-1 -3.i8 - 3.23 -- 4.35
j.06 --n_.r_-1
-4.i5 -4.85
-4.i6 --4.52 -4.53
-4.33 -3.76 -4,iO -5~21
-4.23

- 8.37 - 8.55 - 8.23 - 8.4” -8.69 -8.10 -886 -8.li -8.31 -8.44 -8.61

- 12.06 - 12.46 - 12.18 -13.11 -11.06 -11.18 - 13.44 - 12.64 - 12.61 -11.97
- I 1 4ii

* Cncorrected radiocarbon
ahDERS0i-J (1963). t Uncorrected radiocarbon
ANDERSON (1963).

age = 440 + 150 years. (&Michigan analysis M-1223), KEITH and age = 1890 + 200 years (Michigan analysis M-1224), KEITH and

Carbon and oxygen isotopic composition af mollusk shells

1769

(53) (54) f55f 1561
(57) (53) f59)
Wf (61) P) (63) W-1)
(65)

61-41 61-43 61-47
a-101 6X-102 61-103 61-104 61-106 61-106 6i-107 a-108 6l-IOS 61-X 10 61-111 61-113 61-114, 116,118 6f-117, 119 61-120,1,2
* * * * * *
61-127 63-128,29, 30 61-132-137 61-173 61-175 62-43 6.2-46 62-47 62-45 62-198 62-199 56-636 56-637 56-638
Rivers md streams, Mean

102 Fresh-water samples, mean

* See means of analyses in Table 5.
t Uncorrected mdiocarbon ages: Actlnon.aias 1010 & I50 pare; fMichigsxi sm~1yse.s M-k225 and M-1226), Kzr?xi and ANDEB~ON f1963j.

-4.32 -4.29 --8*52 - 7-37 -7.17 -7.32 -7.22 -7.35 -7.23 -7.35 -7.20 -7.27 -7.40 -7.43 -9~40 -9.31 -**24 - low - f0-07 - IO.15 - 9.33 - 10.24 -9-77 -9% - S-65 -9.97 - lQ”2O -7:85 -6.69 -6.97 -7.39 -7.37 -7*i2 -7*52 - 7.48 -. 3.18 -6*SZ -7*90
--845 --8.46

-11.33 -13.25 - 15.17
-9-51 -9.33 -10.57 -9.81 -11.67 -11.49 - 12-24
--Il*l4
- ltcF?6
_ 12.11
-9.61 -12.35 - 32.52 -12*$4 - 12.53
- 12.52 - 13.07 - 13.07 -11.04 -13.12 -11+53 - 13.83 -1i%s5 -11.51
-9.09 -8,32 - 13.63 --Il.S(i -fl.42 - 13.91 - 12.86 -12‘13 -9.34 -9.88 ---I I *30
--Ii.80 -.
- 9.68

Ligumia 2300 & 200 years

IO

1770

31. L. KEITH, G. 31. ASDERSOS

and R. EICHLER

Table 8. Collection sites of the analysed samples

Location number

Collection site

Depth and
temperature

1

Atlantic at Cape Nay, New Jersey

S*

(4-17”)

2

Mazatlan, W. coast of Mexico

S

(18-31”)

3

Maria Cleophas Isbnd, Tres Marias Islands 60 miles 77;.

of San Bias, Mexico at 21’ 14’ 3.

S

(21-31”)

4

Pacific coast of Mexico, N. of Banderas Bay, 20” 50’ N.,

105” 30’ W.

s

(18-31”)

5

Scammon Lagoon (Ojo de Liebre), Baja California,

Mexico, 27” 45’ N., 114” 15’ W.

1.8

(18-24’)

6

Pacific, off Rio San Lorenzo, Mexico, at 23” 56.2’ X.,

107“ 19.6’W. (Scripps Strt. VS-BII-25)t

53

(14-28”)

7

Pacific, 23” 50.5 N., 107” 18-2’ W. (Scripps Sta. VS-

BII-26)

84

(14-22”)

8

Pa&%, off Rob&r, Sinaloa, Mexico, at 24” 9.0’ N.,

107” 505 W. (Scripps Sta. VS-BII-30)

88

(13-23’)

9

Pa&c, off Puenta Piaxtla, Mexico, at 24” 33.0’ X.,

106” 535’ W. (Scripps Sta. VS-BII-33)

46

(14-27”)

10

Pacific, E. of Tres Marias Islands, Mexico, at 21” 51’ N.,

106” 10’ W. (Scripps Sta. VS-BII-39)

53

(14-28”)

I1

Todos Santos Bay, Punta Banda, Baja California, 50 mi.

5. of Ensenadn, Mexico

31

12

Arbeletos Cove, Punta Banda, Baja California, Mexico

33

13

Pacific, N. W. of Tres Marias Islands, Mexico, at 22” 18’

N., 107” 48’ W. (Scripps Sta. VS-BII-35)

3000

(1.8”)

14

Pacific Ocean at La Jolla, California

s

(16-21”)

15

P&fic, off Balboa Pier, Yewport, California

3,7

( 15-20”)

16§

Beach near Cape Jones, Hudson Bay, Canada, at

54” 30’ N., 79” 20’ W.

S

12-8”)

17

N. E. Hudson Bay, Canada, F. R. l3.1 Sta. 549 (1953),

63” 36’ N., 82”W.

i3

( -0.8’)

18

S. E. Hudson Bay, Canada F. R. B.S Sta. 59-72, at

56” 11’ N., 80” 14’ Wi.

93

( - 1.3”)

19

Hole-in Fog Bay, Ellef Ringnes IsIand, N. W. Terri-

tories, Canada, Lat. approx. 78%.

( - 1 to 05”)

20

Arctic Ocean at Mould Bay, Prince Patrick Island S. W.

Territories, Canada, 76” N., 119’ 45’ W.

25

( -1 to 0.5”)

21

Slide Fjord, near Alert, N. coast Ellesmere Island,

N. W. T., Canada, Lat. approx. 82” PIT.

( -1 to 05”)

228

Hood Canal near Cnion, J&on County, Washington

State

S

(5-25”)

239

Charlton Island, James Bay, N. W. Territories, 70 miles

N. E. of Moosonee, Canada, at 51” 45’ K., 80” W.

s

(0.4 to 10.4”)

24

Shallow bay and stream inlet, Lake Ontario near Jordan,

Ontario

* Depth in meters, temperature in degrees C. S = shallow water biologic community, intertidal to sub-littoral.
t Station of Scripps Vermilion Sea Expedition, 1959. $ F. R. B. = Fisheries Research Board of Cana.da. $ Intermediate salinit,y ranges, 5 to 30 “&.

Carbon and oxygen isotopic composition of mollusk shells

1771

Table 8 cont.

-

Location number

Collection site

25

Piseco Lake, 37 miles N. W. of Amsterdam, Hamilton

County, New York

26

Pembroke Marsh, near HEwmiltonB, ermuda,

27

Pond, Eastern Florid8

28

Houghton Lake, Roscommon County, Michigan

29

Crystal Lake, Benzie County, N. W. Michigan

30

Higgins Lake, Roscommon County, Michig8n

31

Reeds Ltake,Michigan

32

Fond near Thomasburg, H8sting’s County, Ont8rio

34

Paget Msrsh, Paget Parish, Bermuds

35

Kenosee Lake, near Csslyle, S. E. Saskatchewan,Canada

36

Crooked Lake, Ssskatchew8n Cranaxta

37

Midway Lske, N. W. of Lao La Ronge, Sask8tchewan

38

Belcher Islands, Hudson Bay, CctnadaF. R. B.* Sta.

58-34

3Qt

Whitefish Lake, N. W. Territories, C!8nada, 62” 37’ N,

106” 44’ W.

40t

Pell8tt L8ke, N. W. Territories, &r&da, 65” 10’ N,

109” 4O’W.

41

Burt Lake, Cheboygsn County, Michigan

42

Long Lake, Adirondacks, New York

43

Lake Erie, Baach at Evsns Center, S. W. of Buffslo, N. Y.

44

Lake Erie 8t Put-in Bay, Ohio

45

Lake Erie at Athof Springs, S. of Buffalo, N. Y.

46

Lake Ontario 8t Sodus Brty, Monroe County, N. Y.

47

Lake Huron at Bruce Beech, near Kincardine, Ontario

48

Lake Erie at Sterling State Park, near Monroe, Michigan

49

Lake Erie at East Herbour Besch, near %ndusky, Ohio

50

Georgian Bay, Lake Huron, near Midhmd, Ontario

51

Gmnd River, Michig8n at 2 miles E. of Ionic

52

Gmnd River, Michigan, at 1 mile W. of Seranac

53

Silver Spring, Marion County, Florida

54

Wakulla River, near Wakulls Springs, Florid8

55

Shallow ditch ne8r Higgins Lake, Michigan

56

TennesseeRiver 8t mile 303, near Decatur, Al8bsma

57

French Creek 8t Cochranton, Pennsylvsnis

58

French Creek at 2 miles N. of Meadville, PennsylviLnia

59

French Creek at 5 miles N. of Meadvilb, Pennsylvchnia

60

Allegheny River between Warren and Kinzua, Penn-

sylvania

61

Black Creek, Monroe County, New York

62

Stream, inlet to Zukey Lake, Lakeland, Livingston

County, Michigan

63

Little Portage River, Washtenaw, on Washtenew-

Livingston County line, Michigan

64

Meramec River, Meramec State Forest Park, Franklin

County, Missouri

* F. R. B. = Fisheries Research Board of Canada. t Specimensfrom locations 39 and 40 8re from stomach sssnplesof a f&h: Coreganur&r pa-
formis.

1772

M. t. KEITH, G. M. ANI)ERSUN and R. EICHLER

Complete salinity data are not available for most of the collection sites. Samples
included in the marine group are from normal marine waters with salinities in the range of 30 to 36x,. Bata at hand show a salinity range of 34.47& to 35+0%, for Pacific Ocean sites (2) to (13) off the Hexican coast, 33.3x0 to 33.77$$for locations ( 14) and (15) in the near coastal Pacific off southern California, and 30$&t, o 3257& for locations ( 17) to (21) in the Canadian Arctic.
The fresh-water species which were analysed are from relatively shallow lake
and river sites (less than 2 meters deep); no data were obtained regarding the range of water temperatures. Three sites, locations 16, 22 and 23 are in waters of intermcd-
iate salinity, between 5 and 30x,, The shells of modern fresh-water mollusca, including pelecypods (Fig. 1) and
g~tropods (Fig. 2) have oxygen and carbon isotope ratios different from those of marine shells ; the fresh-water shells are relatively deficient in both oxygen- 18 and
carbon- 13.

Within the marine group, the principal differences from one environmental sub-group to another are in the oxygen isotope ratio. It is evident that the differences are in accord with the well-known temperature dependent oxygen isotope fractionation between calcium carbonate and ocean water. The highest oxygen-18 content is found in the shells from deep sea and Arctic Ocean sites, the lowest in the shells
from warm Pacific waters off the coast of Xexico. The fresh-water shell sampfes are ~I*-deficient, relative to the marine samples,
an expectable consequence of the H,tY8: lXJY6 fractionation which takes place during evaporation and precipitation in the weather cycle (EP~TEIXand MAYEDA, 1953). The latitude effect observed by those authors, i.e. a progressive depletion of oxygen-18 in rain and surface waters from higher latitudes, is observable in the fresh-water shell samples, for example, in the sequence from the Meramec and Tennessee River shells (lat. 35’ to 38’N.) to those from the more northerly Grand River and Allegheny River (lat. 42” to 43”N.) A similar latitude-dependent variation of oxygen isotopic composition can be seen in the gastropod shells (Fig. 2); the highest oxygen-18 content is found in the fresh-water shells from Bermuda and Florida, (indicated by “3” and “F” in Fig. 2) and the lowest in gastropod shells from northern Iakes of the Canadian Shield. Two of the specimens from sub-arctic lakes,* marked “A” in Figs. 1 and 2, have extreme oxygen isotopic composition, (60 18= - 15.2 and - 17 per mil), well beyond the scale of the figures.

The analysed marine and fresh-water shells have different ranges of carbon isotopic composition; the marine shells fall between +4*2 and -1.7 per mil,i_ and the fresh-water shells between -2.1 and -15.2 per mil, with two exceptions,
* Samples 61-151, 152, i The compositional range of the marine she& is similar to that reported by C~are (1954). He records one extreme analysis (fKF3 = -3,8 per mil) for a peeten shell specimen from Apafaehee Bay, Florida, a marginal marine environment in which some continental carbon cont,ribution from river waters is to be expected.

Carbon and oxygen isotopic composition of mollusk shells

1773

TWO fresh-water gastropod shells (“G” on Fig. Z), both ~~~0~~~8 from eastern Lake Erie, have carbon isotope compositions which overlap the range of marine shells. H. VANDEE SCHALXEof the University of Michigan reports (personal communication) that Goniobasis feeds on the aquatic plants growing on boulders in shallow water. It is possible that they may ingest some calcium carbonate from the rocks, or more probably, from the carbonate precipitated on some types of aquatic plants, which is relatively enriched in carbon-13 (CRAW, 1953).
Among fresh-water shells, the most striking difference between environmental sub-groups is in the carbon isotope ratio. Shells from rivers (Table 7 and Figs. 1 and 2) are relatively deficient in carbon-13, with a mean KP of -11.9 per mil, in comparison with those from large clear lakes, such as the Great Lakes (mean KY3 = - 4-3 per mil). The shells from brown bog-water lakes * have carbon isotopic compositions which overlap the range of river specimens.
The observed carbon- 13 deficiency of fresh-water shells, and the relative carbon-13 deficiency of river shells in comparison with those from large lakes, probably are due mainly to a variable land plant and humus contribution to fresh-water systems. CY-deficient carbon resulting from carbon fractionation by land plants can be contributed in two principal ways: by the pathway of carbon dioxide from plant respiration and humus decay added to ground water and streams, and by the incorporation of humus particles, soil microbes and derived material in the food web of aquatic animals.
BAERTSCHI(1951), CRAIG (1954) and CLAYTCBaNnd DEGIZNS(1959) referred to the expected effect on fresh-water carbonates of CO%from decaying organic matter. KEITH, EICELER and PARKER (1960) proposed that the food web, rather than dissolved bicarbonate in the water, may be the main direct source of carbon in shell carbonate. The problem of environment-controlled differences in carbon isotope ratios was re-examined by KEITH and ANDERSON (1963), utilizing data on Cl4 content as well as Cl3 content of marine and fresh-water shells; their results support the probability that the land plant and humus reservoir of carbon has an important effect on continental waters and on the carbon isotopic composition of fresh-water shells.
CONSIDERATIONOF POSSIBLE VITAL EFFECTS
A firm conclusion regarding environmental control of carbon isotope ratios of shells requires that we examine an alternative explanation: the possibility that the observed differences may be due primarily to vital effects, differing from one species to another, rather than to environmental differences. The analysed marine and fresh-water specimens are of different genera of course, and most of those from river sites are different genera and species from those collected in lakes, The question is whether different species may fractionate carbon isotopes to different degrees and thus exhibit differences which could erroneously be attributed to environment.
In order to resolve that question, we have made a study of isotopic variability within and between communities of mollusks. A comparison of three mollusk species
* Four open circles clustered around the Allegheny River data {Fig. 1) represent pelecypod shell6 from Precambrian shield lakes of Canada and the Adirondack ares; their carbon isotopic
compositions are similar to those of gastropods (Fig. 2) from small lakes of the Canadian Shield.

1771

3%. L. KEITH, G. N. XXDERSONand R. EICXLEK.

from a river community with the same three species from a lake community* is presented graphically in Fig. 3, In each case? several samples of exterior shell carbonate were analysed, in a progression from older to younger shell, in order to show within-shell variability.
It is evident that the largest and most significant difference between the analysed lake and river shells is environment-controlled rather than species-controlled. The three species from the Grand River in Michigan have shells with a mean Xi3 of

LAKE ERIE

t

t

20

SAMPLE

,

I

40

DISTANCE

t FROM

, 60
BEAK

!

I

80

mm

Fig. 3. Carbon isotopic composition of three pelecypod species from a lake and a A\-cr. Mean KY8 values of exterior shell subsamples of Elliptio, Pusconaiu and Lampsilis are -8.68, -8.80 and -8.31 per mil for specimens from Grand River, Michigan; --7.71, -7.94 and -7+55 per mil for the Lake Erie specimens.

- 12.0 per mil, whereas the same three species from Lake Erie have shells with a mean KY3 of -4.1 per mil, with no overlap between the lacustrine and fluvial shell

data.

CARBON ISOTOPIC COMPOSITIONOF SOFT PARTS

Carbon isotopic composition of the soft parts of invertebrates have previously been reported, for example by CRMG (1953), BROECKER and OLSON (1961), but the data on fresh-water mollusks are very limited and it appeared desirable to compare lake and river specimens with marine specimens. Two marine, three

lacustrine and five fluvial specimens of air dried pelecypod flesh were oxidized to yield carbon dioxide samples following the procedure of CRAIG (1953). Isotopic analyses, compared with carbon dioxide from shells of the same specimens, are presented in Table 1.
Two observations appear to be worthy of mention: firstly, the carbon isotopic composition of pelecypod flesh samples are characteristic, as are the shell compositions, of the environment in which the animals lived. There are three distinctive groups with 6C13 decreasing in sequence from marine to iacustrine to fluvial samples. Secondly, the mean isotopic difference between shell and soft parts is about the same, 17 to 20 per mil, regardless of environment. In view of the limited number of samples,

* Collected with the advice of HENRY VAN DER SCRALIEof the University of Michigan.

Carbon and oxygen isotopic composition of mollusk shells

1775

the greater difference between shell and soft parts of lacustrine specimens is not
regarded as significant. * It is concluded that vital effects are less important than environmenti effects in
controlling the carbon isotopic composition of pelecypod soft pszts and shell.

FIBROUSARAQONITEIN LIGAMENTS
In connection with the study of mollusk shells, an examination was made of the external ligament of several species of marine and fresh-water pelecypods. The ligaments consist of closely-packed parallel fibers embedded in a brown organic matrix. Fibers make up the bulk of the interior portion of the ligaments examined but are absent in the exterior portion and adjacent to the attached shell surfaces. If the brown organic matrix is removed by solution in Clorox (5% solution of sodium h~ochlo~~), the fibers separab from one another; they are silky and flexible, give the X-ray difiaction pattern of aragonite, and dissolve with effervescence in
cold dilute HCI. It is of interest that the ligament fibers are composed of the aragonite form of
calcium carbonate in all specimens examined, regardless of whether the shell also is aragonite, as in Tivela (Table 3) and fresh-water pelecypods (Tables 4 and 6), or whether the shell is wholly or in part composed of calcite, as in Chlamys (No. 6244) or the Ostrea and Mytilw specimens examined (Table 3).
Carbon and oxygen isotopic compositions of ligament fiber samples are given in Tables 3, 4 and 6. With the exception of the oxygen in one M~tiZu.s specimen (&Z-62), the ligament fibers are enriched in both oxygen-18 and carbon 13, in comparison with average shell isotopic composition of the same specimens. The average difference of isotopic composition between shell and ligament (8 specimens) is about 0.5 per mil for 60fe and about 1 per mil for SCP.
Invertebrate ligaments probably deserve further study. They may contribute appreciable quantities of tie-pained acicular aragonite to detrital carbonate sediments. For the present investigation, the significant observation is simply that shell carbonate and ligament fiber carbonate are isotopically different. It follows that the calcium carbonate precipitated by pelecypods probably is laid down in equilibrium with local environments created by the animal, and does not necessarily attain isotopic equilibrium with the external environment. Isotopic differences between exterior and interior shell layers, discussed in the succeeding section, lead to the same conclusion.
W~~-SHE~ VARIATIONOF ISOTOPICCOMPOSZTXON
The variability of carbon and oxygen isotopic composition within single shells was investigated as a necessary basis for choosing a method of sampling and for understanding the differences within and between biologic communities.

Exterior versus interior shell of marine pelecypods
As a tist step, exterior and interior shell samples were analysed (Table 2), beginning with an oyster shell (62-52) in which interior and exterior are calcite, and a

* BXWE;CKESand OLSON(1961) give 6CI9 = -22 per mil for Margaritifera

f&h, &Y =

-4.8 per mif for the shelf, i.e. a shell-fleshdifferenceof 17.2 per mil for a p&qpod f&m the

out.let of Lake Tahoe.

1776

M. L. KEITH, G. M. ANDERSON and R. EICHLER

clam shell (62-132) in which both are aragonite. The remaining data in Table 2 are for specimens in which the exterior layer is calcite and the interior pearly layer is aragonite.
The difference bO~~~n8t-erio6r0t.&,i,, is negative for specimens of Atrina. and Mytilus, positive for single specimens of four other pelecypod species analysed. It would, of course, require more extensive sampling to establish a generalization regarding species-controlled differences.
The difference: 6C1In”terior - E&rior has a positive sign in all of the analysed marine pelecypod shells except one, an oyster (62-52) with complicated shell structure possibly due to extreme changes of environment during it,s growth. Although the exterior and interior layers are calcite, several irregular intermediate layers are composed of aragonite.*
Neither the oxygen nor the carbon isotopic data of Table 2 show any consistent difference which can be correlated with the aragonite or calcite form of shell carbonate.

Exterior shell layers of marine mollusks
A second stage of the investigation of isotopic variability within single shells involved analyses of exterior shell sub-samples in sequences from older to younger shells. For individual marine pelecypods, discussed above, the variation of oxygen and carbon isotopic composition of exterior shell is less than one per mil and the range within any one specimen does not overlap the composition of interior-shell. This is not the case for marine gastropods.
Two large marine gastropod shells were studied in some detail. One of them, Lunatia from 34 meter depth off the Pacific coast of Mexico, shows a relat,ively small changes of isotopic composition in successive samples of exterior shell, possibly due to a nearly constant environment. The exterior shell is relatively deficient in carbon13 (6 = -0*44%,), as compared with the interior pearly shell (6 = T 1.04%,), a difference similar to that observed in marine pelecypod shells.
The other large gastropod, a 9 in. long conch shell (Strombus) from 1.5 meter dept,h in the Florida Keys, shows systematic changes of both oxygen and carbon isotopic composition in a sequence of exterior shell sub-samples ground to less than 1 mm depth from the tips of successive spines. The data are shown graphically in Fig. 4. Oxygen isotopic composition exhibits a cyclical variation which almost certainly is related to seasonal changes of water temperature, and is analogous to the cyclical change observed by UREY et al. (1951) in a belemnite. There are two complete cycles of change of SO18 (Fi g. 4), which leads to the rather surprising conclusion that this large specimen of Strombus gigas was only two years old.7
The conclusion that Strombus grows and deposits shell carbonate at a very rapid rate is supported by the observation of the University of Miami Marine Laboratory.

* 4 separnt,cd amgonite sub-sample gave tTW = -2.63 per mil (appreciably diffvr+wt front
exterior and interior shell) and gave NT13 = --0.12 per mil, intermediate between the carbon isotopic compositions of exterior and interior lnyfxs.
t Previous estimates give a mwh longer life span; for example, ABBOTT (lg.?:;p,. 21) states that large specimens probably rcprrscnt trm to twenty-five years of qro\vth.

Carbon and oxygen isotopic composition af mollusk shells

1777

that tank cultured Strmbus grow f&m the larval stage to a 2 m. length in about

5 weeks (R. JOHNSON and L. GREEN-FIELpDersonal communication).

Xt should be noted that our method of sampling to obtain an ordered sequence of

growth increments of conch shells is entirely different from that adopted by EPSTEIN

and LOWZ~TAM

(19531, who sampled f&m&us shells by cutting out a 5 x 5 mm

rectangular column from the lip shell. They analysed successive layers ground from

that columnas piece, beginning at the inside, and interpreted the resultant sequence

of oxygen isotopic analyses as representing seasonal growth temperatures over a

I

I

1

I

I

I

I

I

IO

30

40

SPIN? NUMBER

Fig. 4. Within-shell isotopic variation of a marine gastropod:

Strambus gigas.

period of years. It now appears that their sampling procedure and interpretation should be applied only to mature specimens. S&Y&W individuals which have
reached maximum size add to the thickness of lip shell. by depositing new layers on tile inside ; however, if specimens grow to a nine-inch length in about two years, then a small segment of lip shell would represent, for most specimens, growth of only a part
of one se:&8on. In that connection, EPSTEIX and ~WENSTAM recognized that the calculated range of about 8°C in shell growth temperature of a Strombu+s specimen from Bermuda was considerably less than the 14’ annual range of water temperatures, By contrast, the range of 2.16 per mil in 601s, which we find by sampling the
exterior spines of a ~~~~~ from the Florida Keys, is equivalent to a shell growth
temperature range of about lO”C, very close to the actual range of water temperatures at our oolleotion site.
The earbon isotope ratio of the exterior spine carbonate changes within a narrower range (Fig, 4). There is an abrupt change of trend at about mid-point in the sequence of spine samples, Le. at the point taken to represent the end of the first year of the animal’s life span. Possibly a migration or change of feeding habits took place at that time.

Exterior shell layers of fresh-water pelecypods
The average range of oxygen-18 content (minimum to maximum) within the external layers of single fresh-water pelecypod shells (Table 4) is only 0.65 per mil,

1776

31. L. KEITH, G. 31. AXDERSONand R. EICHLER

similar to the range of variation in two marine pelecypod shells (Table 3). and probably reflecting a general tendency for pelecypods to deposit shell carbonate only within a limited range of water temperature (cf. EPSTEIN and LOWENSTAM, 1953).
The carbon-13 content of fresh-water pelecypod shells varies over a much wider range ; external shell sub-samples (Table ‘i) show an average within-shell variation of Z-6 per mil in 6C13. The largest within-shell variation of carbon-13 content encountered (5.1 per mil) is that of a fluvial pelecypod, Actinonaias (Table ri and Fig. 5).

-16-
45$ -14WY 3-C1D3 -
-12-
-II -

0
‘\ ‘*_-’

/
/ l’

Fig. 5. \Vithin-shell carbon isotopic variation of a fluvial pelecypod:
CCWi7LChZ.

dctinonaiua

The sequence of change in that shell appears to be quite regular; 6C13 increases slightly at first, and then decreases from -11.31 to - 16.56 per mil in the most recently deposited shell carbonate.
The observed within-shell changes of carbon isotope ratio could be due either to a change of carbon isotopic composition in the local carbon cycle of the river environment or to a change of feeding habits of the pelecypods. Evidence from an examination of ~vithin-commul~ity variation, discussed in a succeeding section, leads one t.o choose the second alternative: a change in feeding habits.

The santpling problem
The largest within-shell variation of 6018 is in a marine gastropod, whereas the largest variation of SLY3 is found in fresh-water pelecypods. The most extreme ranges of variation are such that sampling of a whole shell or a small portion of shell would not influence the isotopic differentiation of marine from fresh-water shells nor the differentiation of lacustrine and fluvial pelecypod shells. Our practice of sampling larger shells by cutting a representat,ive slice across growth bands aud ituAuding bot,h inner and outer shell layers, appears to be a defensible compromise.

Carbon and oxygen isotopiccompositionof mollusk&K&S

1779

However, it is clear that inv~sti~at.ions of within-s~ci~ 01‘ ~t~-~o~u~ty variability can best be made by sampling the entire shell of specimens to be compared, or by careful selection of comparable samples taken from a particular part of
each shell.

W~mm-COMMUMTY

VABU~ON OF SHELLISOTOPICCOMPOSITION

The present work includes studies of three marine and seven fresh-water mollusk ~mm~ti~,* from each of which six or more cohabitant specimens were analysed. It is considered that the other sites listed in Table 7 have not been sampled adequately to permit evaluation of within-community variation. Data reported in Table 7 are for samples obtained by grinding an entire shell of small specimens or a representative cross section of large shells.

The within-community range of 60’* is less than one per mii for shells from each

of the above noted booties,

both marine and fresh-water. The average &G&R-

community range is 0.59 per mii, essentially the same as the average within-A&

range of dOI8for pelecypods. Our data show, on the average, a slight enrichment of

oxygen-l 8 in marine gastropod shells, relative to the shell of cohabitant pe1ecypods.t

The differences are small but they are resent and in the me direction as those

observed by EPSTEIN and LOWENSTAM (1953), who attributed the difference to a

tendency of pelecypods to deposit shell only during the warmer part of the year. The

sampling problem is particularly diEcult for gastropods, in which ~th~-shell

variation may be ~~derably larger (Fig. 4) than the mean d.Serenees between

g~tropo~ and pelecypods.

Our data do not show any pronounced differences in oxygen-18 content among

pelecypod shells of Merent cohabitant species. Detailed study of recently de-

posit& lip shell from river clams of the French Creek oommunity (Table 5), shows

some differencea of dOI*,for example between cohabitant Lam&&s and Pkurobemu,

but the differencea are small and possibly are due to d.iEerenm in the temperature

range over which the animals actively deposit shell carbonate. At a more southerly

site, in the Tennessee River (Lot. 66, Table 71, shells of eleven different qe&s of

pelecypods gave &Ox8between -7.17 and -7.43 per mil, i.e. nearly identical within

the limits of measurement.

i%&%
With-~5mmu~ty variation of SC13, based on analyses of representative cross sections of shells, is muuh larger than that of 60x8. The average within-community range of cRYsfor three marine collection sites (2.45 per mil) is about the same as the average range for seven &&-water sitea (2.21 per mil).
The obsemation that carbon isotopic oomposition may vary widely within a single shell of some fresh-water pelecypods (e.g. Fig. 5) prompted US to modify

* Table 7, marine locations 2, 3 and 14, fre&-water locations 44, 51, 56, 67, 59, 61 and 64. Marine location 1 i not considered hare because the specimens are dead shell from the beeh and rqreaent an unknown range of offshore depths.
t %%kxMi )‘* (inpermil) 0f&wtropod8 @f andpelecypods fPf....Loc. 2: G( -%16), P( -2-71); Lot. 3: G[ -I.%+ P( -%PO); Lo% x4: G( -Q-72), Pf -1.04).

1780

&I. L. KEITH, C. N. ANDERSON and R. EICHLER

our sampling procedure to allow a better comparison from one specimen to anothel.. Data in Table 5 are analyses of rectangular pieces cut from the lip of each shell outside of the palial line, to represent shell laid down during approximately the same period of time.
The data show some differences in carbon isotopic composition from one species to another. For example, lip shell samples of Elliptio, Lasmigona and Lampstlis fuscioEa from the French Creek community (Table 5) have average rXY assays near

0

II

I

/

f

t

/

i

20

40

60

80

100 120

WELGHT OF SINGLE VALVE , GRAMS

Fig. 6. Variation of shell carbon isotopic composition within a community of fresh-water peleeypods: Actinonaim carinata.

-13 per mil, whereas lip shell samples of cohabitant pink and white Pleurobema have an average &Y3 near -11 per mil, with no overlap into the range of 6V3 of the other three species.
The maximum observed within-community differences of KY3 are between small and large individuals of the same species rather than between one species and another. The isotopic composition of a size sequence of Actinonaias from one biologic community* was investigated in some detail. The analytical results (Table 5 and Fig. 6) show a progressive decrease in carbon-13 content of lip shell samples with increasing size of the animal, over a range of KY3 from - 10.5 per mil to - 16.3 per mil. A similar tendency, i.e. for lip shell deposited by large individuals to be depleted in carbon-13 relative to shell concurrently deposited by small individuals, can be observed in data for other species from French Creek (Table 5) and also in the QuadruEa sequence from the Grand River at Dunnville, Ontario. The size sequence of Cumberlandia from the Clinch River is quite different from the others and exhibits an irregular contrary trend, toward increase of KY3 in shell laid down by larger individuals.
It is noteworthy that the range of BY3 of lip shell concurrently deposited by a size sequence of cohabitant Actinonnias individuals (Fi g. 6) is essentially the same as the
* Specimens collected alive from the E’rench Creek, 5 miles north of Jleadvifle, Pa.

Carbon and oxygen isotopic composition of mollusk shells

1781

range of KY3 observed within the exterior layers of a single large shell (Fig. 5). Therefore, the differences cannot be attributed to a change of carbon isotopic composition in the environment. We propose that the variation is probably a consequence of a change of feeding habits, which takes place gradually as the animal gets older. It is known from tank culture of oysters that they reject some food particles and digest others (H. J. TURNER, Woods Hole, personal communication) and a similar selective ability probably is a general attribute of
pelecypods.

CARBONISOTOPICCOMPOSITIOINN RELATIONTO Mo~usx

METABOLISMAND TO SOURCESOF C-BON

It has been shown that there are wide differences of carbon-13 content between

marine and fresh-water mollusks and between lake and river mollusks, and the

differences between environmental groups can be observed by comparing KY3 of

shells, or equally well by comparing KY3 of ligament fiber aragonite or mollusk flesh.

It has been shown that there are systematic differences of &Y3 between exterior and

interior shell layers and between shell and ligament fiber carbonate. In addition,

there are differences of carbon-13 content among the shells of cohabitant mollusks of

different species, and even larger differences between small and large individuals within a biologic community.

It appears from the evidence, as recapitulated above, that the observed broad

scale differences of carbon isotope ratio are controlled mainly by the isotopic

composition of the external environment. Smaller scale its-moll~k

and within-

community differences probably are dependent upon internal environments main-

tained by the mollusk, and can best be explained by the hypothesis that the organic

carbonate is not deposited in carbon-isotopic equilibrium with dissolved bicarbonate

in the water (CRAIG,1953). The present evidence appears to favor an additional or

corollary hypothesis, namely that the carbonate carbon in shell and ligament is

mainly metabolic and is derived from the food web rather than directly from dis-

solved bicarbonate. The land plant and humus reservoirs of carbon affect the bicarbonate, the food

web and the biologic communities of continental waters in several stages, with

consequent CY3-deficiency in those of continental waters relative to those of the

ocean and in those of river waters relative to those of lakes. In a fnst stage, a part of the CO, from terrestrial plant respiration and from oxidation of humus in soils

will be added to ground water and thence to rivers and to lakes. That CO,, with

KY of about -26 per mil, will lower the average carbon-13 content of bicarbonate

in continental waters, more in rivers than in lakes, and may be regarded as opposing

the effect of exchange of atmospheric CO,(KP = -7’ per mil) across the air-water interface (BROECKERand WALTON,1959) as well as the more local effects of disso-

lution of marine limestones (NY3 close to zero) and of fermentative CO, from bottom

muds (OANAand DEEYEY, 1960).

Atmospheric CO,, with KY3 about - 7 per mil, is an important contributor to the carbon cycle in lakes, but probably plays a minor part in the formation of bicarbonate in ground waters contributing to rivers, due to the low CO,-content of the atmosphere (043 ~01% f, too low to account for the solution of carbonates and silicates in the

1782

MM.L. KEITH, G. M. ANDERSONand R. EICHLER

amounts necessary to produce hard ground waters. In contrast with the above figure, soil gases may contain about O-3 vol. y0 of CO, within the upper 10 cm and up to 10% or more at a depth of 1 or 2 meters (VOGEL, 1959).
A second stage, additional to the ground water effect, will result from t.he oxidation of humus within the waters and sediments of rivers and lakes. The net effect of the two stages or paths of addition of land plant carbon will vary according to the nature of vegetation and soils in a drainage basin and may be very small in barren areas. In general the land plant effect and consequent lowering of Cl3 content in bicarbonate and dependent food web will be large in rivers because of large proportional amounts of suspended humus and of ground water,* and smaller in lakes, where much of the humus has been removed by settling and where the long residence time of water permits prolonged exchange with atmospheric CO,.
A third stage humus effect is suggested, based upon the hypothesis that mollusk shell carbonate is mainly derived from metabolic CO,, and therefore is affected by the carbon isotopic composition of the food web and by food selectivity of the mollusks. The change of shell 6C13 of some fluvial pelecypods with age (Fig. 6) is in a direction to be expected if they adapt themselves gradually to utilize a larger proportion of humus and humus-derived materials.? It is to be expected that the food web in humus-laden streams will be less homogeneous than that of large clear lakes. If fluvial pelecypods adapt the&elves to select and digest humus-derived components in preference to bicarbonate-dependent components of the food web, the selective process will accentuate the carbon isotopic differences between pelecypods from lakes and from rivers.
Radiocarbon assays of mollusk shells, discussed briefly in a succeeding section. are consistent with the hypothesis that the Cl* content as well as the Cl3 content of bicarbonate, food web and mollusks, is affected by contributions of carbon from the humus reservoir to continental waters.
The major differences of 6CY3 in mollusk shells are between environmental groups (see Figs. 1,2 and 3) and are clearly environment controlled; the smaller scale differences attributed to metabolism and food selectivity of mollusks do not result in appreciable overlap in the range of carbon isotopic composit,ion of marine, lacustrine and fluvial shells.
ARCHAEOLOCJICALCONSIDERATIONSAND RADIOCARBON DATING
The present work has a bearing on archaeology in two respects. Firstly, it is apparent that measurement of carbon and oxygen isotopic composition of shell fragments or artifacts provides a means of determining whether they came originally from the ocean, from lakes or from rivers (Fig. 1). In some instances it may be possible to go a step further and to differentiate between alternative river sources of
* Most streams are fed mainly by ground water, except during the short run-off period following a rain; in addition, they are continually re-working soil materials from run-off and river bank erosion.
t If the observed change of&Y with age is the result of a change of feeding habits, it would appear that Cumberlandia (Table 5) differs from Actinonaias and from the other fluvial pelecypods studied in size sequence, and does not adapt itself to utilize increasing amounts of hum&. Perhaps there is a difference in the origin and evolutionary adapt.ion of Cumberlandia compared to the other species.

Carbon and oxygen isotopic composition of mollusk shells

1783

shell material. The data show, for example, that one can ~fferentia~ Tennessee
River shells (latitude 34%) from Ahegheny River shells (latitude 42°K). Secondly, measurement of both carbon- 13 and carbon-14 content of modern
shell samples provides some insight regarding the assumptions upon which radiocarbon dating is based. Three river shells, three marine shells and one lacustrine shell,* all from specimens collected alive, were submitted to the Michigan Radiocarbon Laboratory for measurement of radiocarbon age.
Radiocarbon assays reported separately (KEITH and ANDERSON,1963) show a
variation of Cl4 which parallels that of Cl3 and is apparently a dilution effect rather
than a fractionation effect; fresh-water shells are CY4-deficientrelative to marine shells and river shells are CY-deficient relative to lake shells. The three fluvial specimens, with a mean &Y3 of -13.06 per mil gave radiocarbon ages of 1010 to 2300 years, attributed mainly to incorporation of inactive carbon from humus, probably via the food web as well as by the pathway of carbon dioxide from humus
decay. Local Cl4 deficiency of a fresh-water system may be due to incorporation of
inactive carbon from limestones, as suggested by GODWIN(1951). The effect was demonstrated by DEEVEYet al. (1958) and the idea was extended by BROECKERand
WALTON(1959), who attributed variations in the CY4/C12ratio of fresh-water bicarbonate to two factors: (a) the relative amounts of carbonate and silicate dissolved, and (b) the rate of exchange with atmospheric CO,. In view of the present results, including the widespread characteristic difference of NY* between lake shells
and river shells, it appears more logical to attribute the difFereneesof both C1eand Cl4 content mainly to a variable humus cont~bution, in competition with the atmospheric CO, contribution, and to regard dissolving carbonate-bearing rocks as
having a superimposed local effect in drainage basins where limestones are abundant.
On this basis, appreciable errors of radiocarbon ages are to be expected for fresh-water shell samples, whether or not they are from a limestone area. Maximum errors probably will be encountered in dating shells from mollusks which lived in streams which were actively eroding old flood plains or well developed soil profiles.

APPI~~OATI~N TO FOSSILS
The existence of a characteristic range of carbon isotopic composition for modern fresh-water mollusk shells and the conclusion that the difference between them and marine shells is due mainly to a land plant and humus effect in continental waters, leads to the expectation that the difference should be diagnostic in well-preserved mollusoan fossils from rocks formed as far back as the Carboniferous, when land plants had reached a stage of extensive development. That conclusion is m accord with the data of CLAYTONand DEUENS(1959), whose fossil and limestone samples were mainly of Carboniferous age, and in accord with a recently completed extension of their work to samples of different geologic ages (Kxrrn and WEBER, in press).
A schematic representation of depositional environments in relation to variation of carbon and oxygen isotope ratios of pelecypod shells is presented in Fig. 7, as a means of summarizing the various effects which have been discussed. Figure 7 is
*Marine specimens61-155, 62-132, 62-196; Fluvial ep&mens 61-11, 61-127, 62-48;
Lacustrine specimen 61-16 (see Table 7).

1784

M. L. KEITH, G. M. ANDERSOX and R. EICHLEK

partly conjectural but is based mainly on present and previous work, including the

ocean paleotemperature studies at Chicago and elsewhere, the work of EPSTEIN

and MAYEDA

(1953) on oxygen-18 in continental waters and the work of CRAIG

(1953, 1954), CLAYTON and DEGENS (1959) and others on variation in the carbon

isotope ratio. Data on mollusk shells from transitional environments will be pre-

sented in a separate paper (KEITH and PARKER, in preparation). The figure is

presented with overlapping environmental fields and without numerical scales because

r

1

RIYERSANDSMhLLZAKESlNAREAS

WITH ABUNDANT YECETATION AND

SOIL COVER

Low Altitude
or Latitude

High Altitude
or Latitude

LARCLLAKESAND THEIROUTLETS

r’
/

Low Altitude.
ar Latitude
TRANSITIONAL ENYIRONMENTS
Marginal Bays. Dtltis. Estuaries

High Altitude or Latitude

Fig. 7. Idealized

INCREASlNC 016 : 0” RATIO -
relationships between environmmt position of mollusks.

and shell isotopic com-

it is intended to show only directions of change. There are many complicating factors and many exceptions are to be expected.

AcknowZedgenaents-We

are greatIy indebted to HENRY vax DER SCHM.IE (University of Michi-

gan) for samples and for identification of fresh-water mollusca, and to ROBERT PARKER (Scripps

Institute of Oceanography) for identification of most, of the marine specimens. J. N. WEBER

gave able assistance in the later stages of the laboratory work and measured the isotopic

composition of the soft parts of several specimens. T. H. VAN ANDEL kindly arranged for the

senior author to do some collecting and sample preparation at the Scripps Institution of Ocean-

ography and t.o participate in the Vermilion Sea Expedition of 1959. Various specimens were

supplied by P. A. BUTLER and 11’. T. YOUNG (U.S. Fish and )Vildlife Serv.), A. H. CLARKE

(Nat. Museum of Canada), E. H. GRMNGER and L. G. MCMULTON (Fisheries Res. Board, Canada),

Lmz COZAD and R. R. LANKFORD (Scripps), D. S. Rawso~ and F. M. ATTOX (Univ. of Sas-

katchewan), E. J. ROSCOE and ALAN SOLEX (Chicago _Museum Sat. Hist.), and several others.

including SHEILA COWAN, ANDREW and DAVID KEITH, DAVID XANELLA and H. G. SK.%vLE%f.
L. F. HERZOG and T. Esmw supervised construction of the mass spectrometer.

Carbon and oxygen isotopic composition of mollusk shells

1785

The investigation was supported by the Earth Sciences Division of the National Science Foundation (grants G-9388 and G-18947) and a supplementary grant was received from the Research Comittee of the American Association of Petroleum Geologists.

REFERENCES

ABBOTT R. T. (1953) American SeasheUs. Van Nostrand, New York. B~~nrsonr P. (1951) Relative abundances of oxygen and carbon isotopes in carbonate rocks.

Nature, Lond. 168, 288.

Booarnn 0. B. (1936) The shell structure of mollusks. K. Dawk F’iidensk., K. Skr. Nate.

viderwk. 2, 232.

BROECXER

W. S. and OLSON E. A. (1959) Lamont radiocarbon me~urements VI. Amer. J.

SC&, RudiocuTbon Sup& 1, 1I.1.

BROECKER

W. S. and OLSONE. A. (1961) Lamont radiocarbon measurements VIII. Rudio-

carbon 3, 176.

BROECKER

W. S. and WALTONA. (1959) Geochemistry of Cl* in fresh-watersystems, (feochim.

et Cosnzochim. Acta 16, 15.

&AX-TOS R. N. and EPSTEINS. (1958) The relationshipbetween Ois/Or* ratios in coexisting

quartz, carbonate and iron oxides from various geological deposits. J. Qeot. 88,362.

CLAYTOWR. N. and DEOENSE. T. (1959) Use of carbon isotope analyses of carbonates for

differentiating fresh-water and marine sediments. Amer. Assoc. Petroleum 17eol. Bull.

43,890.

CLAYTON

R. N. (1961) Oxygen isotope fractionation between calcium carbonate and water.

J. Chem. Phya. 34, 724.

CRAIG H. (1953) The geochemistry of the stable carbon isotopes. return. et Concern.

Acta 3, 53.

CRAIGH. (1954) Carbon-12 in plants and the relationshipsbetween carbon-13 and carbon-14

variations in nature. J. Qeol. $2, 115.

CRAIGH. (1957) Isotopic standards for carbon and oxygen and correction factors for mass-

spectrometric analysis of carbon dioxide. Geochim.et Cosmoehim. Acta B, 133.

DEEVEY E. S., GROSSM. S., HUT~EINSONG. E. and KRaysnt H. L. (19543 The natural Cl4

contents of materials from hard water lakes. Nat. Ad. Sci. Proc., Wad&g&n D.C. 40,285.

EICELER R. (1961) Carbon and oxygen isotope ratios in marine and fresh-water mollusk shehs.

M.S. Thesis of the Department of Geophysics and Geochemistry, the Pennsylvania State

University.

EPSTEINS., BUC~SBAUMR., LO~ENSTAMH. A. and UREY H. C. (1953) Revised oarbonate-

water isotopic temperature ecale. Bu.& Qeol. Sot. Amer. 84, 1315.

E~~TJZINS. and LO~EHSTAMH. A. (1953) Temperature-shell-growthrelations of Recent snd

~~rgl~i~ Pleistocene shoal-water biota from Bermuda. J. Qeol. 81, 424.

EPSTEINS. and MAYEDA T. (1953) Variation of Ois content of water from natural sources.

Qeochim. et Cosmochim. Acta 4, 213.

GOODWINH. (1951) Comments on radiocarbon dating for samples from the British Isles. AmeT.

J. Sci. 249, 301.

GRAP D. L. (1960) Geochemistry of carbonate sediments and sedimentary carbonate rocks,

Part IV A: Isotopic composition and chemical analysis. I&no& State QwQti Sway,

Circular 308.

JEFFE~YP. M., CO?.ZFSTOWN., GREENEALGRD. and DE LASTERJ. (1955) On the carbon-13

abundance of limestones and coals. (3eochim. et Co8mochim. Acta 7, 255.

KEITH M. L. and ANDERSONG. M. (1963) Radiocarbon dating: fictitious results with mollusk

shells.Science141, 634.

KEITH M. L., EXCELJZRR. and PAIXER R. H. (1960) Carbon and oxygen isotope ratios in marine

and fresh-water mollusk shells (abstract). Bull. GeotlS. ot. Amer. 71, 1901.

KEITH M. L. and WEBER J. N. (1964) Carbon and oxygen isotopic composition of selected limestonesand fossils. Qeochim. et Cosmochim.Acta $?& 1787-1816.

MCCREAJ. M. (1950) On the isotopic chemistry of carbonates and a paleotemperaturescale.

J. Chem. Phys. 18, 849.

11

1786

BI. L. KEITH, G. M. ANDERSONand R. EICHLER

_&KINNEY C. R., MCCREAJ. M., EPSTEINS., ALLENH. A. and UREY H. C. (1950) Improvement,s in mass spectrometers for memurement of small differences in isotope abundance ratios.
Rev. Sci. Instrum. 21, 724. OANA S. and DEEVEY E. S. (1960) Carbon-13 in lake waters and its possible bearing on paleo-
limnology. Amer. J. Sci., Bradley Vol. 285A, 253. REVELLE R. and FAIRBRIDGER. (1957) Carbonates and carbon dioxide. Chapter in Treat,ise
on Marine Ecology and Paleoecology. Geol. Sot. Amer., &lem. 67, Vol. 1, 239. UREY H. C. (1947) The thermodynamic properties of isotopic substances. Amer. Chem. Sot.
J. 562. CREY H. C., LOWENSTAB~H. A., EPSTEIN S. and MCKINNEY C. R. (1951) Measurement of
paleotemperatures and temperatures of the Upper Cretaceous of England, Denmark, and the southeastern United States. Geol. Sot. America Bull. 62, 399. VOGEL J. C. (1959) Eber den isotopengehalt des kohlenstoffs in siisswasser-kalkablagerungen. Geochim. et Cosmochim. Acta 16, 236. WILBUR K. M. (1960) Shell structure and mineralization in mollusks, in Calcification in Biological Systems (Ed, R. F. SOGONNAES)A. mer. -45~0~. Adv. Science 64, 15. WEBER J. N. and KEITH M. L. (1963) Isotopic composition and environmental classification of selected limestones and fossils (abstract). Geol. Sot. ilmer.. Special Paper 73, 259.