Geochimrca n Cosmochimrca ACIP Vol. 48. pp. 625-539 Q ~ergamon Press Ltd. 1984. Printed in U.S.A.

0016-7037/84,53.00

+ .OO

Nitrogen and carbon isotopic composition of bone collagen from marine and terrestrial animals
MARGARET J. SCHOENINGER* and MICHAEL J. DENIRO** Departmentof Earth and Space Sciences,Universityof California,Los Angeles, California 90024

(Received May 1I, 1983; accepted in revised form December 14, 1983) Abstract-The stable nitrogen and carbon isotope ratios of bone collagen prepared from more than 100 animals representing 66 species of birds, fish, and mammals are presented. The b15Nvalues of bone collagen from animals that fed exclusively in the marine environment are, on average, 9% more positive than those from animals that fed exclusively in the terrestrial environment; ranges for the two groups overlap by less than I Wo. Bone collagen 6N values also serve to separate marine fish from the small number of freshwater fish we analyzed. The bone collagen bN values of birds and fish that spent part of their life cycles feeding in the marine environment and part in the freshwater environment are intermediate between those of animals that fed exclusively in one or the other system. Further, animals that fed at successive trophic levels in the marine and terrestrialenvironment are separated, on average, by a 3% difference in the 6N values of their bone collagen. Specifically, carnivorous and herbivorous terrestrialanimals have mean b15N values for bone collagen of +8.0 and +5.3%0,respectively. Among marine animals, those that fed on fish have a mean b15Nvalue for bone collagen of + 16.X& whereas those that fed on invertebrates have a mean 615Nvalue of +13.3%0. These results support previous suggestions of a 3%~enrichment in 615N values at each successively higher trophic level. In contrast to the results for dN values, the ranges of bone collagen 6C values from marine and terrestrial feeders overlap to a great extent. Additionally, bone collagen 6C values do not reflect the trophic levels at which the animals fed. These results indicate that bone collagen 6N values will be useful in determining relative dependence on marine and terrestrial food sources and in investigatingtrophic level relationshipsamong differentanimal species within an ecosystem. This approach should be applicable to animals represented by prehistoric or fossilized bone in which collagen is przserved. INTRODUCTION
PREVIOUSWORK suggests that the stable nitrogen and carbon isotope ratios of marine plants and animals are different from those of terrestrial plants and animals (DEUSER et al., 1968; DEGENS et al., 1968; WADA et al., 1975; SWEENEY et al., 1978; WADA, 1980, TAUBER, 1981; CHISHOLM et al., 1982). If this is so, animals feeding on marine organisms should have isotopic compositions characteristically different from those feeding on terrestrial organisms because the b15N and 613C values of an animals tissues are determined by the corresponding isotope ratios of its diet (DENIRO andEpsrH~, 1978, 1981; BENDER~~al., 1981; MACKO ef al., 1982). Furthermore, the isotopic composition of organisms feeding in either environment should also depend upon their position in the food chain. This follows from suggestions that there is a trophic level effect on the 6N and iC values of animals (MIYAKE and WADA, 1976; PANG 1967; WADA and HATTORI,

and 6C values of animal tissues are, respectively, about 3k and lk more positive than those of the animals diets (DENIRO and EPSTEIN, 1978, 198 1; BENDER et al.. 1981; MACKO ef al., 1982). These observations are consistent with the existence of the proposed trophic level effect for nitrogen isotopes and a small trophic level effect for carbon isotopes as well. Nevertheless, the hypothesis that marine and terrestrial organisms have different 15N/14Nand 13C/C ratios is not strongly supported by the available field data. Although the mean of published 6N values for land plants is 4k less positive than the mean for marine plants, the ranges overlap completely (Fig. 1). The same is true for the S13Cvalues in plants. The range of values for terrestrial plants overlaps completely the range for marine plants (DEINES, 1980). In addition, the dN values of terrestrial animals are overlapped by those for marine animals (Fig. 1). Likewise, in carbon there is considerable overlap between the d&ibutions of 6C values of marine and te animals
(DEGENS, 1969; SCHWARZ, 1969).

and NRIAGU, 1977; MCCONNAUGHEY and MCROY, 1979; RAU et al., in press). According to these workers, b15N and 613Cvalues should become more positive as nitrogen and carbon are transferred along the continuum from plants, to herbivores, to primary carnivores, and finally to secondary carnivores. Laboratory experiments indicate that the average 6N * Permanent address: Department of Cell Biology and Anatomy, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205. l * Also Archaeology Program.

The existence of trophic level effects on the b15N and 6C values of organisms is suggested, although not strongly supported, by the available field data. Marine and freshwater zooplankton appear to have d15N values that are, on average, 3160more positive than associated phytoplankton (MIYAKE and WADA, 1967; WADA and HAITORI, 1976; PANG and NWGU, 1977), but there are few data (Fig. 1). Results for animals that feed at higher trophic levels also sugecestthe existence of an enrichment in 15Nin successive trophic levels (HOERING, 1955; MIYAKE and WADA, 1967;

625

676
tvkmne

M. .i
Fish

Schoeninger

and

M.

J. LkNiro

___e-.
. .

-2

MOIIUSCS
Zooplankton

hlorme

Marme

Marme Algae. Phytoplankton, Seaweed Marine Blue-green Algae Terrestrlol Animals

Terrestrial Plants NZFwers

. fi

. 1U:J

i il t

II

) 11 -50

1)

11 00

11

11 11 11 11 +I00 +50 85NA,RWoo0)

11

11 11 +150

11

1 +200

FIG. 1. Published 6N values of whole organic matter of plants and animals from the marine and
terrestrial system. Only plants and animals collected in their natural habitats are included. Data are from BARDIN ef al., 1977; DELWICHE and STEYN, 1970, DELWICHE ef al., 1979; DENIROand EPSTEIN, 1981; HOERING,1955; MIYAKEand WADA, 1967; PANG and NRIAOU, 1977; PARWEL et al.. 1957; RENNIE er al., 1976;STEELE and DANIEL,1978; VIRGINIA and DELWICHE, 1982; WADA, 1980; WADA and HATTORI, 1976

RAU, 198 I), but there are even fewer data available for organisms at these higher levels (Fig I). In the case of carbon, VOGEL (cited in VAN DER MERWE, 1982) found no difference in 613C values of the same type of tissues from animals feeding at different trophic levels in a terrestrial environment, whereas FRY et al. (in press) and MCCONNAUGHEY and McRov ( 1979) observed small increases in the % content when they compared animals living at successively higher trophic levels in the open ocean. The d13C values of open ocean species from successive trophic levels presented by RAU et al. (in press) exhibit the same trend, whereas no consistent trend is obvious in their data for coastal water species. In this paper, we report the results of an investigation into the effects of marine versus terrestrial feeding and of trophic level on the bN and 6t3C values of animals. Trophic levels within food webs overlap to a large extent (COHEN, 1978) and the most common approach has been to reconstruct feeding strategies of individuals within discrete food webs (MEARNS, 1982). In this project, on the other hand, we analyzed more than one hundred modem fish, birds, and mammals from multiple food webs to determine if their isotopic compositions would reflect e&&s of marine versus terresuial feeding and of trophic level in a context broader than individual food webs. We analyzed bone collagen because its isotopic composition re.flects that of the animals diet (DENIRO

and EPSTEIN, 1978, 1981; BENDER et al., 1981), its chemistry is well defined, and its amino acid composition varies only slightly between species (VEIS, 1964). Thus, differences in the bN and 6r3C values of collagen from different animals should reflect differences in the isotopic composition of their diet and not differences in chemical composition. Additionatiy, the choice of bone collagen allows analysis of the remains of animals from prehistoric and earlier periods if the bones still contain collagen. Thus, the results of this investigation have important implications for biologists and ecologists interested in vertebrate feeding relationships, for soil chemists and biochemists interested in nitrogen and carbon cycling, and for prehistorians interested in estimating subsistence strategies of prehistoric humans.
MATERIALS AND METHODS

The animals inch&d in this study and their diets are iisted in Table 1. These animals represent a wide variety of marine
and tenwrkl vertebratesand were chosen so that more than onetrophickveIismpreaentedineachsystem.Bothterrcstriat herbivores and tw&rial camivores were selected, akh0u8h as is true of their rekive abundance in the wild, far fewer carnivores than herbivores were avaiIable for anaIysis. Although only one marine herbivore was sampled, the ether

InarineanimslslhesekctaireplwenttwobmadtrophicLevck
The first group will be. referred to as primary carnivores in the remajnde.rof the paper. In the context of this project, these are vertebrate that feed on invertebrates including

Bone collagen isotopic composition

627

Table 1. Diet, colbction site, and bone collagen 613C and 615N values for specimens analyzed in this study.
GSN(O/oo)

bup
Marine MAMMALS Order: Camivora

Scientific

and Gwnmon Namer

Dier

Site-

SC(e/oo)

423. Zulophus californianus Sea lion 424. i!alophua californianus

fi fi fi fi fi
fi fi fi
m0

SC SC NC NC SC SC NWE NWE AK AK NC NC NC

-12.9 -13.3 -11.0 -13.3 -11.7 -13.9 -11.2 -15.2 -14.1 -13.9 -13.6 -10.0 -9.7 -9.6

+17.6 +17.2 +19.3 +23.0 +16.0 +19.3 +17.5 +15.6 +12.1 +12.5 +12.4 +13.9 +13.3 +15.4

178. Zalophua califomianua See lion 179. Zalophur californianus Sea lion 428. Phoca uitulina Ha&or seal 429. Phoca uitdina
Harbor seal

430. Phoca uitulina Ha&r seal 133. Pagophilus groenlandicua Harp aeal 134. odobenus rosmarus W&US 425. Odobenus rosmams
WdNI

mo

426. odobenus rosmarus W.&W 445. Enhydra lutrir

mo

mo

mo

Sea otter 447. Enhudra lubir

46 1. Delphinopterus leucar
White whale

fi fi fi fi c fi fi fi fi fi fi fi fi fi fi fi fi

CA AK SC SC SC SC SC SC SC SC SC SC SC SC SC SC SC

-13.9 -14.2 -11.9 -12.3 -12.0 -13.4 -14.9 -13.3 -11.6 -13.3 -12.7 -12.1 -11.9 -12.6 -13.5 -12.6 -11.6

+15.9 +17.8 +16.9 +16.3 +15.8 +15.5 +16.5 +14.7 +17.2 +15.2 +14.3 +16.0 +15.1 +16.5 +14.8 +15.6 +16.6

462. Delphinagterua leucas White whale 450. Clobicephala macrorhynchur

pilot whale

451. Clobicephah macrorhphus ShotGinned pilot whak 452. Globicephola macrorhynchur Short&noed pilot whale 431. Delphinus delphis Common dolphin 432. Delphinus delphis
Common ddphin 433. Delphinur delghir Common dolphin

434. Tuniopr truncalw Bottle noH dolphin 436. Tursiops truncatua Bottle nose ddphin 437. Lupenorhynchw obliquidens Pacifii white-sided dolphin 439. Lagenorhynchus obliquidenr
Pacific white-sided dolphin 440. Lopenorhynchur obliquidenr

Pacific white-sided ddphin 44 1. Phocoenoides dalli DidIs porpoise 442. Phocoenoider dalli Dells potpoise 444. Phocoenoides dalli Dalis porpoise 443. Phocoena phocoena Ha&r porpoise

628
Table Group 1. Continued

M. J. Schoeninger and M. J. DeN~ro

Scientific and Common Namer

Diet*

SiteU

GJC(O/oo)

WN(%o)

Marine MAMMALS Order: Cetacea Suborder: My&&i 175. Bofoenoptera pbgsai~ Fin whale 457. Bolaenoptera acutorostroto Minke whale 458. Bolaenoptero ocutorostrata
459. Minke whale Eafaeno#Hera muscuius Blue whale pl,ar,mo pl,ar.mo pLar.mo pt,ar,mo

NC SC SC SC SC
AK AK AK SC SC

-15.4 -15.4 -12.6 -14.5 -12.6 -16.1 -16.4 -15.3 -13.6 -13.0

+12.7 +13.8 t14.9 +11.7 +lS.l? +14.4 +15.4 +13.8 +l4.0 +12.0

460. Botoenoptera muscufus Blue whale 453. Boloena mysticetus Bowhead whak 455. f3nfaena myticetus Bowhead whak 456. Baheno mysticetos Bowhcad whak 448. Eschrichtius robustus Gray whak 449. Eichrichtius robustus Gray whale Marine FISH 159. Scorpoenichthys mormorotus Calxzon 160. Kotsuwonus pefomis Tuna 161. Gengonemus finite White croaker 163. hfenticirrhus undulotus California corbina 164. Citbarichthgs xonthostigma Longfin satxidah 165. Jfoiocentms ~~~bjt~~i~ Squirrel fish 166. Strong&m exilis Needlefish 167. forichtus notatus P&fin midshipman 168. Uyperprosopon argenteum Walkye surfperch 170. Bogre panomensis Catfish Reef FISH 353. Calomus sp.
%JY

pl.ar.mo pl,ar,mo pl,ar,mo pl,ar,mo pl,ar.mo pl,ar,mo

mo,ar fi mo,ar mo,ar mo,ar mo,ar fi mo,ar mo,ar ii

SC SC SC SC SC SC SC SC SC SC

-10.0 -14.2 -12.1 -11.9 -12.7 -13.6 -11.1 -14.4 -12.4 -13.0

+14.9 +11.4 +12.5 +14.1 +14.8 +14.4 +16.0 +13.7 +14.7 +11.1

mo,ar pf,ar
m0

BA BA BA BA

-8.2 -4.9 -4.5 -7.8

+6.6 +6.5 +!i.o +3.9

354. Hff~ufon

sp.

Grunt
356. Albulo uufpes Bone fish 355. Sporisoma sp. Parrot fish Marine BIRDS 157. Ardeo herodins Great blue heron 158. Pelecanus occidentolis Brown pelican 174. Pekcanus occidentolis Brown peIican 217. Fubnclrus sp. Southern fulmer 218. Fregata sp. Frigate bird 2 16. Aptenodvtes sp. King pang&n 220. Eodgptes sp. Macaroni penguin ii fi fi mo,ar fi,mo fi,mo mo.ar
NC SC SC FA YU

al

-14.0 -13.6 -14.5 -19.6 -12.1 -19.0 -18.1

+15.7 +17.0 +17.9 ill.9 +12.4 +13.0 +9.4

FA FA

Bone collagen isotopic composition

629

GouP

Scimtific

and Canmm

Namar

DieP

Site FA FA FA

~C(O/Oo)
-14.9 -15.6 -17.9

GN(/oo) +10.9 +10.8 +10.3

Marine BIRDS 22 1. Pygocelis rp. Gent00 penguin 222. Pygocelir sp. Cent00 penguin 223. Pygocefis sp. centoo penguin Anadromour FISH 146. Oncorhychus nerka Pink safmon 147. Oncorhynchur tahowytscha King &non Aquatic Migratory BIRDS 172. Podiceps nigricolfis Earred grehe 17 1. Aechmorphorus occidentalis Western grebe 169. Nycticorer nyticorox Black crown night heron Fresh-water FISH 144. lctalurus Catfish 145. lctalurus Catfish 148. lctalurus Catfish 143. Ictafurus
nebulosus

ma,=
mo,ar mo,ar

ar fi

AK AK

-17.2 -19.3

+9.6 +12.7

ar fi fi

SC SC SC

-26.3 -18.6 -20.9

+9.1 +14.2 +10.1

de,fi de,fi de,fi de,fi fi

MI TN MX FL MI

-12.7 -23.6 -19.6 -23.7 -19.1

+8.6 +6.8 +8.3 +6.6 +9.5

furcatus furcalud

nebulosus Catfish 149. Saluelinua namaycush Lake trout

Terrestrial BIRDS 219. Chloephaga picla

FA SC SC SC

-22.5 -12.6 -16.9 -21.2

+6.0 +4.9 +6.8 +9.7

Terre&al

MAMMALS

Barn owl 150. Dipodomys rp. Kangaroo rat 151. Sciurus sp. Squirrel 156. Citellus rp. Squirrel 153. Neotoma lepida Wood rat 155. Neotoma lepida Wood rat 154.lmpus rp. Rabbit 699. Loma sp. Uama 700. Lama sp. Lfama 701. Lclmo sp. Lfama 702. Lama rp. Llama 703. Lam0 rp. Llama 251. Oois arier Sheep 463. Ovis aries Lfunb 252. Equus rp. Horae
P P P PO PO PO PS PI PS P P P P P

SC SC SC SC SC SC PE PE PE PE PE SY SC SY

-12.4 -19.9 -19.0 -11.9 -16.9 -22.0 -20.6 -20.5 -20.5 -20.2 -20.5 -20.5 -20.5 -21.2

+8.4 +2.1 +4.9 +2.4 +9.2 +1.9 +5.8 +6.0 +6.5 +6.6 +4.1 +4.7 +5.4 +3.4

630 Table 1. Continued

M. J. Schoeninger and M. J. DeNIro

-~-------_____-. &P Terre&al MAMMALS 253. Capm rp. Goat 283. Copra sp. Goat 254. Bos tourus cow 263. Sus scrofa Pie 477. odocoilour hemionur Deer 627. Leo tigrir Tiier 631. Alopex lapopus Arctic fox 633. Felis concofor Mountainlion 636. &IX canadmrir Bobcat PS PI PI
PS

_- ..^ GW(/oo) -20.3 -17.9 -20.6 -14.0 WN(o/,) +6.9 +7.3 +3.4 +5.0 +5.8 +8.1 +5.9 +7.6 +lO.O

Sciitific

and Cormnan Namer

DietL

SiteSY CY SY

PI ma ma ma ma,bi

FL IN AK SC SC

-22.0 -18.0 -20.6 -17.6 -15.8

l Dieb are abbreviatedas followa: al, algae; ar, arttuopob; bi, birds;de, detritus;fi, firh;p. grain:ma, mammalr;mo. molti

pl, pbnkton; p. plants; ro, rodents.

Taken from: BARNHART 1936; MURPHY, 1936; PALMER, 1982; RANDALL, 1968; FITCH, 1969; BAILEY etal., 1970.; FITCH and LAVENBERG, 1975; WALKER, 1975; THRESHER, 1980; SCHREIBER, 1982. Ster are abbreviatal a. fdbwr: AK, Al&a; BA, Bμ CA, Canada, CY, Cypns; FA, Falkland lrhnb; FL, Rotida; IN, India;MI, Michigan;MX, Mexico; NC, NorthernCalifornia; NWE, NorthwertemEurope;PE, Peru;SC, Southem Calfiomia:SY, !$+a; TN, Tennessee; YU, Yucatan

plankton, arthropods, and moUuscs. The animals in the higher broadly defined trophic level will be referred to as secondary carnivores. These are vertebrates that feed on other vertebrates. Bone samples were cleaned mechanically before collagen was extracted. In the case of very fresh bone, dermestid beetles were used to remove most of the flesh; remaining traces were removed using dissecting scissors and scalpel. We took care to analyze only those museum specimens that had been cleaned in the same fashion. The bones were dried, then powdered to less than 0.7 1 mm. Collagen was prepared from the bone powder as described previously (DENIRO and EPSTEIN, 1981). One modification greatly simplified collagen preparation. All chemical treatments were performed in a funnel fitted with a coarse fritted-glass filter and attached to a tube which had an affixed stopcock. All chemical treatments involved in collagen preparation were done using this ap paratus in conjunction with a filter flask. In this manner, sample transfers were avoided and the risk of losing portions of samples was eliminated. Collagen samples were combusted using a modified version of the STUMP and FRAZER (1973) method (NORTHFELT et al.. 1981). The resulting CO2 and N2 were separated and purified in a vacuum system by cryogenic distillation prior IOdetermination of their isotope ratios by mass spectrometry. The isotope ratios are reported in the 6 notation, where

values, and - I. 1 f 0.2%0for 815Nvalues. Analysis of collagen samples prepared from two aliquots of bone powder from each of twelve samples from this and other studies (SCHOENINGER et al., 1983; DENIRO and SCHOENINGER, 1983) produced means and standard deviations (1 SD. values) of the differences between each pair of 0. I f 0.2% for 6j3C values and 0.2 + 0.3% for li15Nvalues.

RESULTS Collagen 6N values and 613C values are listed in Table I. The 6N values and 613C values of bone collagen from all the animals we sampled are plotted in Fig. 2 and Fig. 4 respectively. In these figures, the animals are grouped into the following categories: terrestrial mammals and terrestrial birds (both groups having fed exclusively in the terrestrial system); freshwater fish (lake or river dwellers that fed on detritus when small and on other fish when large); aquatic migratory birds (birds that spent part of the year along the marine coast and the rest of the year on inland lakes); anadromous fish (fish that spent part of their life cycle in the ocean and part in freshwater streams); marine fish, marine birds, and marine mammals (all having fed exclusively in the marine system); and reef fish (from shallow water flats off a Bahamian island). Means and standard deviations (I S.D. V&KS) and ranges of the 615N and 813C values for these different groups of animals are presented in Table 2. These statistics are presented in order to illustrate the variation within each group. No statistical test for difference

1 =vle _,1 x1m960. &I+ = (15W4N)

(WW

sample _

,m960

( 13C/zC)standard

(15N/14N)standard

The standard for 6C measurements is the Peedee belemnite (PDB) carbonate, while that for dN measurements is atmospheric (AIR) nitrogen. The means and standard deviations (1 S.D. values) for 27 analyses of a thiourea standard were -23. I -t 0.3%0 for bC

Bone collagen isotopic composition I

631

.Hr

= rl&.lr

I
.

I :.
n

Marine ^. Uircts

mm

Anadramous Fish Aquattc Migratory Birds

,:

yI prcciucm

of

lyl+rJplc analyrlr

r.. I .,:

Fresh-water Fish Terrestrial Birds

r 0 = carnivores -. 8 . w..w...I 0. . 0 ;,I

Terrestrtal Mammals

l = herbwares

II

11

11

11

,,,,,,,,,,,,,

FIG. 2. Bone collagen 6N values. Each point represents a sample from one individual. All animals were collected in their natural habitats. For the indicated groups, the means are given by the central vertical

bars, the standard deviations by the heavy horizontal bars, and the ranges by the narrow horizontal bars. The trophic levels occupied by individuals are indicated only for terrestrial animals.

between means was perfomxd because neither the samples that we analyzed nor the populations from which they were drawn can be assumed to be normally distributed. For example, the population of all terrestrial mammals is most likely skewed because there are more species of herbivores than of carnivores. In addition, our samples do not represent randomly drawn samples from all possible populations of any group that is included. Such a sampling strategy was beyond the scope of this project. The results shown in Fig. 2 and in Tables 1 and 2 indicate that there is a discrete difference between the 6N values of bone collagen from animals feeding in the terrestrial system and those feeding in the marine system. For 4 1 marine mammals, the lowest 6r5N value is + 11.7% Marine fish (n = IO) have 615N values that range from + 11.4 to + 16.0!%, and marine birds (n = 10) have bN values that range from +9.4 to + 17.9% In sum, the br5N values of animals feeding exclusively in the marine environment range from +9.4 to +23.0%0 with a mean 6i5N value of + 14.8 + 2.5% (n = 61). Among the terrestrial animals (including mammals and birds), on the other hand, the highest 615N value is +10.0%0 and the mean is +5.9 & 2.2% (n = 27). There is only a 1950overlap between the ranges for those feeding in the marine and in the terrestrial system; and the mean for marine animals is

almost 9% more positive than the mean for terrestrial animals. If only mammals are considered, the range of 6i5N values for terrestrial mammals is separated from the range for marine mammals by about 2960 and the means are separated by almost 10% Although we recognize that the mean P5N value for any category is affected by the numbers and types of animals that were selected to make up the category. these differences in means and virtual lack of overlap in ranges of 615N values in such a large variety of animals indicate that a real difference exists in the b15N values of bone collagen from marine and terrestrial vertebrates. The one exception to this statement is provided by the reef fish from the Bahamas (Fig. 2). The mean d15N value for these four fish is +5.5 + 1.3960,which is very close to the mean for terrestrial animals. The basis for the low b15N values in these fish will be discussed below. Three groups of animals we studied do not fit into the marine/terrestrial classification. The first of these groups includes five freshwater fish, whose mean bone collagen 615N value is +8.0 + 1.2% These fish are carnivorous throughout most of their life cycles (G. SMITH, pexs. commun.), and it is interesting to note that their bone collagen 6r5N values are the same as those for the six terrestrial carnivores (6i5N = +8.0 f 1.6%~). The other two groups of animals that do not fit into the marine/terrestrial classification are the

632

M. J. Schoeninger and

M. J. &Nil-o

Table 2. Means, standard deviations (2 f S.D. value), and ranges of 6taC and 6tsN vaiues of bone collagen from various groupings of animals discussed in the text. -_I ~~.._~~~~~~____~__ _ _.. ._ USN (0,oo) i% 2 S.D. Terreatkl MAMMALS (~23)
TerrestrialBIRDS (n=4) Terrestrini BIRDS and MAMMALS (1~27)
hWbk0rer CilmivOter

63C (O/00)

Range
+1.9, +10.0 +4.9. +9.7 +1.9. +10.0 +1.9, +9.2

j7 2

S. D.

Raw-w
-22.0, -1l.Q -22.5. -12.6 -22.5, -11.9 -22.0, -11.9 -21.3, -15.8 -23.7, -12.7 -26.3, -19.3, -8.2, -14.4, -14.4, -14.2, -19.6, -14.9, -19.6, -16.4, -16.4, -15.2, -19.6, -18.6 -17.2 -4.5 -10.0 -10.0 -11.1 -12.1 -18.1 -12.1 -9.6 -9.6 -11.0 -9.6

+5.7 + 2.2 +6.8 + 2.0 +5.9 + 2.2 is.3 + 1.9

-18.8 + 2.9 -18.3 + 4.5 -18.8 + 3.1 -18.9 + 2.3 -18.3 + 2.1 -19.7 + 4.5 -21.9 + 3.9 -18.2 -6.3 + 1.9 -12.5 + 1.4 -12.2 + 1.4 -12.6 -15.9 + 2.5 -16.6 + 1.6 -15.5 -13.1 -13.5 -12.8 -13.5 + + + + t 3.1 1.6 2.2 1.1 2.1

Fnshwakr FISH (n=S) Aquaticmigratory BIRDS (n=3) AnadromowFISH (n=2) ReefFISH (I&) Mstine FISH (n=lO)
immtebrate eaters (I&)

+8.0 + 1.6 +8.0 + 1.2 +ll.l +11.1 +5.5 +13.8 +14.1 +13.7 +12.9 +10.3 +14.6 +15.6 t 2.7 f 1.3 + 1.6 + 0.9 ?: 2.9 + 0.7 + 2.5 + 2.2

+5.9. + 10.0 +6,6, +9.5 +Q.l, +14.2 +9.6, +12.7 +3.9, +6.6 +11.4, +16.0 +12.5, +14.9 +11.4, +16.0 +9.4, +17.9 +9.4, +11.9, +11.7, +11.7. +14.3. +9.4. +lO.Q +17.9 +22.9 +15.8 +23.0 +23.0

firh eaten (n=Z) MarineBIRDS (n=lO) invertebratee&era (na) fish eaten (n-6) Marine MAMMALS (~41) invertebmtetatrn (n=l6) fish eaters (~~25) Matine FISH, BIRDS

+I31 t 1.3 +16.7 + 1.8 +14.s + 2.5 +13.3 + 1.7 +l6.5 + 1.9

and MAMMALS(~61)
invertebrate eatam fishenters +9.4, +15.8 +11.4, +23.0 -13.7 + 2.4 -12.9 + 1.2 -16.4, -9.6 -19.6, -11.0

aquatic migratory birds and the anadromous fish. The bone collagen 6N values for both groups overlap the marine and freshwater distributions, retkcting their use of food sources in both environments. In addition to the discrete distributions of bone collagen bN values for animals feeding exclusively on marine or on terrestrial food sources, there is evidence for atrophic level effect in each environment. As shown in Fig. 2, carnivorous terrestrial animals (2 birds, 4 rn~rn~s) have a mean bone collagen 6N value of +8.0 k f .6%, while herbivorous terrestrial animals (2 birds, 9 mammals) have a mean 615N value of +5.3 4 1.9% Within the marine system a 3% difference between broadly defined trophic levels is also observed in our samples, as shown in Fig. 3. Among marine birds, those. that fead on invertebrates (molluscs and arthropods) have a mean 6r5N value of + IO.3 + 0.7% (n = 4) compared with vahms of +14.6 k 2.5% (n = 6) for those that include !kh in their diet. The same difference is observed within marine mammals. Those marine mammals (baken whales, sea otters, and walruses) that feed on mollusg arthropods, and plankton have a mean bone co&en Br5Nvalue of + f 3.6 + 1.3% (n = 161, whereas those (toothed whales, seals, and sea lions) that include 6sh in their diet have a mean PN value of + 16.7 + 1.8%~(n = 25). When all marine animals are considered, fish eaters (n = 30; 2 fish, 3

birds, 25 mammals) have a mean 615N value off 16.5 4 1.9%0. whereas those animals that do not feed on fish (n = 26: 6 fish, 4 birds, 16 mammals) have a mean 6N value of + 13.3 t i .7k. In all of the examples cited above, a difference of 3%~ is observed between broadly defined trophic levels. Results shown in Fig. 4 and in Tables 1 and 2 indicate that bone collagen di3C values also separate animals that fed exclusively in the ocean from those that fed only on terrestrial foods, although the ranges are not as discrete as they are for dN values. For example, although the mean bone collagen SC value for marine feeders (n = 61) is 5.5460less negative than that for terrestrial feeders (n = 27), there is an 8%0 overlap in the ranges of d13Cvalues for the two groups. Similar overlaps in the ranges of bone collagen S3C values are observed when terrestrial mammals or birds are compared with marine animals or birds, although in each case the average 613C value for the marine group is less negative than that for the terrestrial group. Bone collagen 6C values for freshwater fish, for the two groups of animals that fed in both the marine and freshwater systems (aquatic migratory birds and anadromous fish) and for marine birds, have ranges that are virtually indistinguishable from those of terrestrial animals. For those animals that live at least part time in freshwater the reason may be uptake of

Bone collagen isotopic composition

633

Marine Mammals

0 Marine Fish
Fish

A Marine Birds
Precwm of sotoplc onolyos

Arthropods, Moliuscs, Plankton

11

-25.0

-20.0

11

-15.0

-10.0

-5.0

8 13 c&&%0)

Fish

Arthropods, Molluscs, Plonkton

FIG. 3. 6N and a13C values of bone collagen from marine animals plotted accordingto their f&ing preferences. The meansaregiven by the centralverticalbaq the standard deviationsby the heavy horizontal bars, and the ranges by the narrowhorizontal bars. Those animals feeding on invertebrates(plankton, molluscs, and arthropods) are separatedfrom those fzeding on fish.

terrestrial carbon by freshwater organisms (RAW, 1980; SCHELL, 1983). Since very little terrestrial carbon is transferred above the microbial level in the marine system (!SCHELL, 1983), however, input of terrestrial carbon probably does not account for the lack of difference in 1313C values that we observed between marine birds and terrestrial animals. The four reef fish from the Bahamas have bone collagen PC values that are less negative than those in any other animals included in this study (Fig. 4). The average fY3C value of this group is almost 7%0 less negative than the average for marine mammals (see Table 2) and the two ranges are separated by 1%o. As mentioned above, these fish had unique bone collagen b15N values as well. The reasons for their ai3C values will be discussed below. Our results exhibit essentially no trophic level effect on bone collagen 6C values (Figs. 3 and 4). In each case in which animals at successive trophic levels can be compared (terrestrial herbivores and carnivores; primary and secondary carnivores among marine fish, marine birds, or marine mammals), there is a small

(< 1960) increase in bone collagen bs3C values going up the food chain, but the ranges for the two trophic levels overlap almost completely. DISCUSSION AND CONCL,USIONS Nitrogen isotopic composition of bone collagen Our results indicate that the ZiN values of bone collagen from most marine animals an more positive than those for terre&al animals. Additionally, we observed that within the different orders of animals we studied (birds, fish, and mammals), the 6N values of bone collagen reflect particular aspects of the feeding regimes. For example, for the 10 marine birds we studied, the PN values ranged from +9.4 to +17.9% Macaroni and Gentoo penguins, which feed on molluscs and arthropods (MURPHY, i936), have the lowest values (< + 11%). Brown pelicans and a great blue heron, which feed almost exclusively on 6sh (PALMER,1962; SCHREIBER, 1982) have the highest values (> + 15%0).

634

M. J. Schoeninger and

M. J. DeNlro

Marine Mammals Marine Frsh Reef Ftsh Marine Birds Anadmmous F,sh Aquatic Migratory Birds Fresh-wter Fish Terrestrial Bwds Twrestrial ~mmais

8 I--+
I

88
A

; = cornwares . .

P.
l

00

IIrlil

-250

IIllilllllrlll~rl,l,J
-200

= herbivores

-150 813CpDe(%o>

-100

-50

FIG. 4. Bone collagen 6C values. Each point represents a sample from one individual. For the indicated groups, the means are given by the central vertical barn,,the standard deviations by the heavy bars, and the rauges by the narrow horizontal bars. The trophic levels occupied by individuals are indicated only for terrestrial animals. small and later they feed on other fish (G. SMITH, pers. Between these two extremes fall the b15N values for commun.). The relatively high GLSN values (+ 12.7966) the southern fulmar, frigate bird, and king penguin. reflect this emphasis on marine foods. Pink salmon, These three species feed on both fish and inve~ebmtes on the other hand, have a two to three year life cycle (PALMER, 1962). of which only one year is spent in the ocean. They Further, bone collagen 6N values of the aquatic feed on freshwater insects when small and include migratory birds and the anadromous fish appear to be marine arthropods in their diets when large (G. SMITH, influenced both by the mixed freshwater/marine diets pers. commun.). The lower 6N value (+9.6%0) reflects these animals consume and by the position in the food the shorter time in the ocean and their lower trophic chain each species occupies. Over the course of a year, level when compared with the king salmon. the migratory birds we studied eat both marine and In sum, the distribution of bone co&en QISNvalues freshwater fish and invertebrates since they winter along among marine birds, aquatic migratory birds, and the coast of southern California and then disperse inanadromous fish reflects each animals dependence on land for the rest of the year. The 615N values of bone marine and freshwater fbods and also the position of collagen for the 3 species we studied range from +9.1 each species in the food chain. The latter serves as to + 14.2%~. spanning the ranges for pure freshwater further evidence for a trophic level effect on nitrogen and pure marine feeders. The earn!!! g&e, whose diet is composed largely of aquatic and land insects (PAL- isotope ratios of bone collagen. There is one group of animals whose bone collagen MER, 1%2), has the lowest value. The Western grebe, whose diet is composed largely of fish (~80%) (PALMER, 615N values are not consistent with their dependence on marine food. This is the set of four reef fish from 1962), has the highest value. The black crown night the Bahamas. Their &lSN values vary from i-3.9 to heron, whose diet is about half fish and haff arthropods +6.596s, which span the middle of the range for ter(PALMER, 19621, fails in between the two extremes. restrial mammals and are lower than those for other The two anadromous fish we studied are a king marine fish. We believe the explanation for the low salmon and a pink salmon. The king salmon has a b15N values for these fish lies in the large amount of four to five year life cycle of which three years are nitrogen fixation by blue-green algae (STEWART, 1978) spent in the ocean. They feed on arthropods when

gone collagen isotopic composition that occurs in coral reefs relative to that which occurs in other parts of the ocean. CAPONEand CARPENTER ( 1982) reported that the rate of nitrogen fixation in reef systems is 25 g/m-year compared to 0.1 g/myear for the upper 100 m. of the rest of the oceans. Blue-green algae and zooplankton feeding on them have lower bi5N values than phytoplankton which do not fix nitrogen and their associated zooplankton (WADA and HATTORI,1976; WADA, 1980). If the bulk of nitrogen brought into a reef system is that which is fixed by blue-green algae, the relatively low 615N values would be carried up the food chain, thus accounting for the low b15N values observed for reef fish. A detailed study of the i5N distribution in blue-green algae, in the other primary producers, and in the consumers within a reef ecosystem will be required to confirm the explanation we advance here. Some of our results suggest that bone collagen 615N values may also be affected by metabolic processes. The most positive 6i5N values among terrestrial herbivores were observed for a kangaroo rat (6N = +8.4%) and a wood rat (6N = +9.2%0). The kangaroo rat seldom, if ever, drinks water, being able to synthesize water from its dry food. Because of this, kangaroo rats have restricted or delayed nitrogen excretion (SCHMIDT-NIELSEN, 1964; VAUGHN, 1972). Some other rats are able to exist on dry food without water, and are likely to have nitrogen excretion similar to that of the kangaroo rat. This ability varies between populations and species and appears to depend on a particular animals environment (VAUGHN, 1972). It is possible that these unique metabolic adaptations to xerophytic environments may have some effect on the nitrogen isotope fractionation which occurs during amino acid synthesis (GAEBLERet al., 1966). Perhaps amino acid synthesis in non-drinking desert-adapted animals involves isotopic fractionations different from those occurring in other animals, so that more 6i5N is incorporated into collagen. It would be interesting to know in this regard if the second wood rat we analyzed, whose bone colIagen 615N was +2.4%, came from a less xerophytic environment than the wood rat whose b15N value was +9.2%. Unfortunately this information is not available. A summary of our results on the 6i5N values of bone collagen in animals at different trophic levels in the marine and terrestrial systems is shown in Fig. 5. In a simplistic representation of the food chain, we have plotted the means and standard deviations of published values for marine and terrestrial plants at the bottom of the graph. As can be seen from the figure, the standard deviations of the marine and the terrestrial plants overlap. This overlap does not occur between primary carnivores in the two systems. It ap pears that the small difference in mean 6i5N values between marine and terrestrial plants is increased as nitrogen moves up the food chains in the two environments. This magnification may be due to selective feeding by animals. In addition, there is at least one more trophic level in the marine system than in the

635

terrestrial system. In combination, these phenomena appear to account for the observed 9%~ difference in mean b15N values between marine and terrestrial animals even though the means of marine and terrestrial plants are separated by only 4%. The data shown in Fig. 5 also are consistent with the existence of a 3% enrichment in 6i5N values as nitrogen is transferred between trophic levels. The slopes of the lines drawn through the data for the terrestrial and marine systems indicate enrichments of 2.7% and 3.3% per trophic level respectively. Carbon isotopic composition of bone collagen The d13Cvalues of bone collagen from marine animals are less negative than those from terrestrial animals that we studied. The mean 6C values we observed (Table 2) are close to the values that CHISHOLM et al. ( 1982) predicted should be observed for pure marine feeders (- 13%) and pure terrestrial feeders (-20%). The one exception to this observation, the marine birds, will be discussed below. The ranges of 613C values we obtained for marine or terrestrial feeders, however, are larger than have been reported previously. Whereas previous results indicate that bone collagen b13C values for human pop ulations whose diets were primarily marine or primarily terrestrial do not overlap (TAUBER, 1981; CHISHOLM et al., 1982; SCHOENINGER et al.. 1983), the range of 6i3C values we obtained for terrestrial mammals overlapped the range of 6)C values in marine mammals by about 4.5L (Table 2). This observation possibly can be explained by considering the diets of the terrestrial animals whose bone collagen 813C values are in the zone of overlap (- 16.5 to - 12.0%). These include a domestic turkey, a kangaroo rat, a wood rat, a pig, and a lynx. The herbivores may have fed on significant amounts of C, plants, since all feed on grass seeds (WALKER, 1975) and many grasses are C, plants. The lynx may have fed on herbivores whose diets were composed largely of C, plants. The b13C values of C, plants are less negative than those of C, plants (SMITH and EPSTEIN, 197 1; BENDER,197 1). Eating substantial amounts of Cd plants (in the case of herbivores) or animals that have eaten large amounts of C, plants (in the case of carnivores) would produce bone collagen 6C values in the range we observed for the animals in question (DENIRO and EPSTEIN, 1978). There are two observations which suggest that there may be some metabolic difference in carbon isotope fractionation between birds and mammals. First, birds that fed exclusively on marine foods have bone collagen 613C values that are more negative than those in fish and mammals that also fed on marine foods (Fig 4). Second, the aquatic migratory birds that fed in both the marine and terrestrial system have bone collagen 6C values that are more negative than those in terrestrial mammals (Fig 4). Based solely on their feeding regime, we expected that they would have had collagen 613C values intermediate between those for terrestrial

636

M J. Schoenmger and M. J. DeNlro

TERRESTRIAL
Secondary Carnivores

MARINE

Primary Carnivores

Herbivores

II

Illll
+ 5.0

Ill
+100

Ill

II

11

+15.0

+x).0

65N,,R(%o)
FIG. 5. Means and standard deviations of 6N values of plants (published values) and of bone collagen in animals (this study) in the terrestrialand marine environments. The means are indicated by the central points and the standard deviation by the horizontal bars. Plants at the base of the food chains in both systems are shown at the bottom of the graph. The animals are divided into broad trophic levels. The lines are drawn to connect the mean bN values for the organisms at the top and bottom of the food chains in each environment. Plants which fix atmospheric nitrogen and animals feeding on them are not included.

animals and marine animals. The number of samples Diet estimation bused on 61SNand 6% values q/ is too small, however, to take these two observations bone collagen as any mote than a suggestion that birds may process carbon isotopes differently than mammals do. AddiBased on the results of this study, several types of tionally, other explanations may account for some as- diet estimation using isotope ratios of bone collagen pects of these observations. For example, the effect of appear to be feasible. temperature on the d13C values of plankton (RAU et The 6N values of bone collagen can be used to al., 1982) may explain the observation that the six distinguish some animals that fed exclusively on marine marine birds with most negative 6C values were colfoods from others that fed exclusively on terrestrial lected in the Falkland Islands. foods. In our sample, animals with bone collagen 615N The extreme 13C enrichment observed in the four values less than +9% fed only on terrestrial foods; reef fish (Fig. 4) probably reflects feeding in an area those with bone collagen 615N values greater than where seagrasses and coral form a significant part of + 15760 fed only on marine foods. In cases where the the base of the food chain (D. MILLER,pets. commun.; bone collagen d15N values fall between +9 and + 1Sk, other information on living habits must be used in MEYERet al., 1983). A similar enrichment in tissues conjunction with the isotope data in order to identify of fish from coral reefs and seagrass meadows compared with offshore fish has been reported (FRY ef al.. 1982 food sources. Two mammals and one bird out of 25 in our sample of terrestrial animals have bone collagen and in press). The basis for this difference is probably b15N values in this range and, thus, could not have related to the enrichment of 13Cin shallow water reef building corals (LANDet al.. 1975; SWART. 1983) and been identified as animals who fed strictly on terrestrial foods. Of these three, two are carnivores and one is in seagrasses (BENEDICTet al., 1980). an animal that lived in a xerophytic environment. A summary of our results on the b13Cvalues of bone Morphological analysis would clarify two of the three collagen in animals at different trophic levels in the cases since camivory in mammals is often reflected marine and terrestrial systems is shown in Fig. 6. In contrast to the situation for bN values (Fig. 5) d13C by the shape of the posterior teeth in the mouth. Among the marine animals in our sample, almost 50% had values of bone collagen do not appear to reflect, to 615N values between +9 and + 16% and thus could any useful degree, the trophic levels at which animals not be identified as having fed on marine foods based fed in either environment.

Bone collagen isotopic composition

637

TERRESTRIAL
Secondary Carnivores

MARINE

+7;_$),

Primory Carnivores

(n=6)

Herbivores

Lb=
h=21) ~__________------*n~T_&-__----______,

Plants -(n=l148)

-30.0

11

11

11 -25.0

11

11 - 20.0 63cpD*(%o)

-15.0

I
-100

FIG. 6. Means and standard deviations of 6C values of plants (DEINES,1980) and of bone collagen in animals (this study) in the terrestrial and marine environments. The means are indicated by the central points and the standard deviation by the horizontal bats. Plants at the base of the food chains in both systems are shown at the bottom of the graph. The animals are divided into broad trophic levels, as in Fig. 5. The four reef fish are not included for masons explained in the text.

solely on bone collagen PN values. Among mammals and tish this would not pose a problem. Since all known mammals and fish that live in the ocean also feed in the ocean, it would be obvious that low 615N values in animals identified as marine by morphological criteria are influenced by their trophic levels. Among birds, overall body form could be used in conjunction with bone collagen 615N values in order to distinguish between carnivorous birds (e.g., hawks) that feed on terrestrial animals and marine birds (e.g., penguins) that feed on marine invertebrates. In cases in which the animals are known to use both marine and terrestrial food sources, it should be possible to use 615N values of bone collagen to estimate general proportions (although not exact percentages) of dependence on marine and terrestrial foods. The PC values of bone collagen can be used for estimating dependence on marine and terrestrial foods only when there is no possibility of feeding on C, plants. In areas where C, and C, plants are utilized as food sources, those animals feeding on a mixture of marine and terrestrial foods could not be distinguished from those feeding exclusively on terrestrial foods. In geographical areas or within animal populations for which C, plants are known to be unimportant (VOGEL, 1978; VAN DER MERWE, 1982), b13C values of bone collagen are useful in estimating dependence on marine versus terrestrial food sources (TAUBER, 1981; CHISHOLM ef al., 1982; SCHOENINGERet al.. 1983).

In the case of vertebrates living in tropical areas surrounded by reef systems, an exceptional combination of bone collagen PC values (greater than -9460) and bone collagen 615Nvalues (less than +79&o) should serve as a signature of total dependence on marine foods. As far as is now known, this combination of low 6t5N values and high 613Cvalues is unique among animals. Vertebrates feeding solely in the terrestrial system and those feeding in both the marine and terrestrial systems in these areas must be analyzed before further predictions can be made. A vertebrates position in a trophic system can be predicted, within limits, for animals feeding only in the marine system or only in the terrestrial system based on bone collagen P5N values. Within the marine system, animals in our sample with bone collagen 6i5N values less than +13L fed on invertebrates whereas those with P5N values greater than +16.5%0 fed on other vertebrates. More terrestrial animals should be analyzed before a similar prediction can be made for the terrestrial system. It would be difficult, however, to distinguish between vertebrates feeding on both marine and terrestrial foods and those feeding only on marine invertebrates based solely on bt5N values of bone collagen. Other lines of evidence would have to be used if such a situation were a possibility. Our observation that a13C values of bone collagen can not be used in predicting feeding positions in trophic systems may be a result of our sampling strategy. We

638

M. J. Schoeninger and M. J. DeNiro

analyzed individuals from multiple food webs, whereas special thanks to D. McIntyre for her help in obtamng sampies, providing information on diets of marine mammals. previous workers reporting a trophic level effect in and in providing us with the founders of our dermestid beetle carbon analyzed individuals from single food webs colony. We appreciate discussions with B. Gelvin, California (MCCONNAUGHEY and McRou, 1979: RAU et al.. in State University at Northridge. D. Gelvin of Santa Monica press). College, C. Harter of Malibu, and G. Rau, NASA Ames Field which improved the paper and figures.The figureswere drawn The types of dietary analysis discussed above for by J. Guenther of UCLA. Typing was done by T. Sahm and modem animals should also be useful for reconstructB. Coe of the Johns Hopkins University School of Medicine ing the diets of vertebrates from prehistoric and earlier and Gen Kurtin and Cathy Elkins of UCLA. Support for the periods. In such cases, however, postmortem alteration project was provided by NSF grants BNS 79-24756and ATM ?9-2458 1. of bone collagen is a possibility that must be considered. Methods for ascertaining whether or not such alteration REFERENCES has occurred must be developed and applied before 615N or b13C values of bone collagen of prehistoric or BARNHART P. S. (I 936) Marine Fishes qfSouthem Cahftimia. paleontologic age can be accepted as representing the University of California Press. BAILEYR. M., FITCH J. E.. HERALDE. S., LACHNERE. A., values that were present when the animal was alive. LINDSEYC. C.. ROBBINSC. R. and Scot-r W. B. (1970) With this caveat in mind, we have shown that bone A List of Common and Scientific Names of Fishes from collagen d15N values are useful in estimating marine the United States and Canada (3rd Edition). American and terrestrial components of diet in prehistoric huFisheries Sot. #6. mans (!SCHOENINGER et al.. 1983). BARDIN R., DOMENACHA. M., CHEALAMET A. and PACHIAUDI C. (1977)Rapports isotopiques naturels de Iazote A final comment is in order. The sampling strategy Il.-Application a la mesum de la fixation symbiotique de we employed in this study was designed to obtain large Iazote in situ. Rev. Eeol. Biol. Sol 14, 395-402. numbers of species of animals feeding in the marine BENDERM. M. (1971) Variations in the lsCc/isC ratios of and terrestrial environments. Accordingly, we do not plants in relation to the pathway of photosynthetic carbon dioxide fixation, Phytochemistry 10, l239- 1244. have results that allow us to determine with great cerD. A. and STEVETON R. L. (1981) tainty the variability of bone collagen 6N and 613C BENDERM. M., BAERREIS Further light on carbon isotopes and Hopewell agriculture. values among different individuals of a species that Amer. Antiq. 46, 344-353. fed on the same diet. Nevertheless, analysis of the 22 BENEDICT C. R., WONG W. W. L. and WONGJ. H. H. ( 1980) Fractionation of the stable isotopes of inorganic carbon by cases in which we have at least two individuals of the seagmsses. Plant Physiol. 65, 5 12-5 17. same species indicates that the average range of values CAPONED. G. and CARPENTER E. J. (1982) Nitrogen fixation among individuals of the same species is 1.8 rf: 1.2Ym in the marine environment. Science 217, IMO-I 142. for 615N values and 1.9 + 2.4%0 for b13Cvalues. These CHISHOLM B. S., NELSOND. E. and SCHWARCZ H. P. (1982) ranges are about the same as those we obtained from Stable-carbon isotope ratios as a meaSure of marine versus terrestrial protein in ancient diets. science 216, I I3 l-l 132. determination of collagen bC and 615N values from COHEN J. (1978) Food We6s and Niche Space. Princeton the bones of 15 mink raised on the same diet (DENIRO University Press. and SCHOENINGER, 1983). This amount of variability DEGENSE. T. (1969) Biogeochemistry of stable carbon isoamong individuals is small relative to the differences topes. In Organic Geochemistry (eds. G. EGLINTON and M. T. J. MURPHY), Chap. 12, pp. 304-329. Springer. we observed for marine versus terrestrial feeding, but DEGENS E. T., BEHRENDT M., GOI-~HARDT B. and is of the same magnitude as the differences between REPPMANNE. (1968) Metabolic fractionation of carbon 6j5N values of organisms feeding on successive trophic isotopes in marine plankton--II. Data on samples colIected levels. Thus, it is likely that in order to differentiate off the coasts of Peru and Ecuador. Deep Sea Res. 15, 1 I trophic levels based on 6N values of bone collagen, 20. DEINES P. ( 1980) The isotopic composition of reduced organic it will be necessary to analyze several individuals from carbon. In Handbook of Environmental lsolope Geochemeach species being studied. On the other hand, this istry (eds. P. FRITZ and J. CH. FONT), Chap. 9, pp. 329amount of variability in bone collagen 615N values is 406. Elsevier. small enough so that it should be possible to identify DELWICHE C. C. and STEYN P. L. (1970) Nitrogen isotope fractionation in soils and microbial reactions. Environ. Sci. marine versus terrestrial feeding preferences in indiTechnol. 4, 929-935. vidual animals. DELWICHE C. C., ZINKEP. J., JOHNSON C. M. and VIRGINIA R. A. (1979) Nitrogen isotope.distribution as a presumRtive
Acknowledgments-Bone

samples were provided by J. Northern, D. Buth, K. Gelvln, E. San&fur, and W. Keegan of UCLA; G. Smith of the University of Michigan; G. Foster of The Johns Hopkins University Hospital; J. Lipps of the University of California at Davis, S. Drieschman of Seaworld in LaJolla, California, F. John of Los Angeles; D. McIntyre of the Los An&s County Museum of Natural History; E. Wing of the Florida State Museum in Gainesville; J. Wheeler of the Peabody Foundation for Archaeology in Andover, Massachusetts; and H. Tauber of the Carbon- 14 Dating Laboratory, National Museum, Copenhagen. Assistance in preparing collagen samples and determining their isotopic compositions was provided by K. Katrak, M. Minagawa, G. Maat, K. Cracker, P. Farnsworth, H. Ajie and D. Winter. We give

indicator of nitrogen fixation. Bot. Car. 140 (suPP~.). 565569. DENIRO M. J. and EPSTEINS. (1978) In&nce of diet on the distribution of carbon isotopes in animals. Geochim. Cosmochim. Acta 42,495-X%. DENIRO M. J. and EPSTEINS. (1981) InIIuence of diet on the distribution of nitrogen isotopesin animals. Geochim. Cosmochim. Acta 45, 34 I-35 I. DEN~ROM. J. and SCHOENIFK~ER M. J. (1983) Stable carbon and nitrogen isotope ratios of bone colla@at: variations within individuals, between sexes, and within populations raised on monotonous diets. J. Archaeol. Sci. 10(3), 199203. DEUSERW. G., DEGENSE. T. and GUILLARDR. R. L. (1968)

Bone collagen isotopic composition Carbon isotope relationships between plankton and sea water. Geochim. et Cosmochim. Acta 32, 657. FITCH J.E. (1969) Offshore Fishes of Califonria. State of California Resource Agency, Department of Fii and Game. FITCHJ. E. and LAVENBERG R. J. (1975) Tidepooland Nearshore Fishes of Ca&xnia. University of California Pmss. FRY B.,LuTE~ R., NORTHAM Wand PARKER P. (1982) A iQ2C comparison of food webs in Caribbean sea8ras.s meadows and coral reefs. Aquatic Botany 14,389-398. FRY B., SCANL,AN R. S. and PARKERP. (1983) 3C/2Cratios in marine food webs of the Tortes Strait, Australia. Atut. J. Mar. Freshwater Rex in press. GAEBLER0. H., Vtrn T. G. and VUKMIROVICH R. (1966) Isotope effects in metabohsm of N and 15Nfrom unlabeled dietmy proteins. Can. J. Biochem. 44, 1249-1257. HOERINGT. ( 1955) Variations of nitro8en- 15 abundance in naturally occurring substances. Science 122, 1233-34. LANDL. L., LANGJ. C. and SMITHB. N. (1975) Preliminary observations on the carbon isotopic composition of some reef coral tissues and symbiotic zooxanthehae. Limnol. Oceanog. 20.283-287. MACKOS. A., LEE W. V. and PARKERP. L. ( 1982) Nitrogen and carbon isotope fractionation by two species of marine amphipods: laboratory and field studies. J. Exp. Mar. Biof. Ecol. 63, 45-49. MCCONNAUGHEY T. and McRov C. P. (1979) Food-web structure and the fractionation of carbon isotopes in the Bering sea. Mar. Biol. 53,257-262. MEARNSA. J. (1982) Assigning trophic levels to marine animals. In Coastal Water Research Project, biennial Report for the years 1981-1982 (ed. W. BASCOM) pp. 125-141. MEYER J. L., SCHULTEE. T. and HELFMANG. S. (1983) Fish schools: an asset to corals. Science 220, 1047-1049. MIYAKEY. and WADA E. (1967) The abundance ratios of 15NJN in marine environments. Rec. Oceangr. Works Jpn. 9, 37-53. MURPHY R. C. (1936) Oceanic Birds of South America. MacMil~n Co. NORTH~ELT D. W., DENIRO M. J. and EPSTEINS. (1981) Hydrogen and carbon isotopic ratios of the cellulose nitrate and saponifiable lipid fractions prepared from annual growth rings of a California Redwood. Geochim. Cosmochim. Acta 45, 1895- 1898. PALMERR. S. (1962) Handbook of North American Birds. Yale University Press. PANG P. C. and NRIAGU J. 0. (1977) Isotopic variations of the nitrogen in Lake Superior. Geochim. Cosmochim. Acta 41.81 l-814. PARWEL A., RYHAGER. and WICKMANF. E. (1957) Natural variations in the relative abundances of the nitrogen isotopes. Geochim. Cosmochim. Acfa 11, 165-170. RANDALLJ. E. (1968) Caribbean Reef Fishes. T.F.H. Pub. Inc. RAU G. H. (1980fCatbon-13Jcarbon-12 variation in subalpine lake aquatic im Food source implications. Can. J. Fish. Aquatic Sci. 37, 742-746. RAU G. H. (198 I) Low sN/N in hydrothermal vent animals ecoio8ical implications. Nnture 289, 484-485. RAU G. H., Swk?EM!Y R. E. and UPLAN I. R. (1982) Plankton ~c:~ ratio chantles with latitude: differences between northern and southern oceans. Deep Sea Rex 29, 1035. 1039. RAU G. H., MEARNS A. J., YOUNG D. R., OLSON R. J., SCHAFJZR H. A. and KAPLAN I. R. (1983) Animal 13CJ2C

639

correlates with trophic level in pelagic food webs. Ecology. in press. RENNIED. A., PAUL E. A. and JOHNS L. E. (1976) Natural nitrogen-15 abundance of soil and plant samples. Can. J. Soit Sci. 56, 43-50. SCHELLD. M. ( 1983) Carbon- 13 and carbon- 14 abundances in Alaskan aquatic o~isms: Delayed production from peat in amtic food webs. Science 219, 1068-1071. SCHMIDT-NIELSEN K. (1964) Animal Physiology. PrenticeHall Inc. SCHOENINGER M. J., DENIRO M. J. and TAUBERH. (1983) 15NJN ratios of bone col&en reflect marine and termsnial components of prehistoric human diet. Science 220,138 l1383. SCHREIBER R. W. (1982) A study of the Brown Pelican. Nafuraf History 91, 38-42. SCHWARCZ H. P. (1969) The stable isotopes of carbon. In Handbook ofGeochemistry (ed. K. H. WEDEPOHL) vol. 2, Chap. 6B, pp. l-16. Springer. SMITH B. N. and EPSTEINS. (1971) Two categories of 13CJ l2c ratios of higher plants. Plant Physiol. 47, 380-384. STEELEK. W. and DANIEL R. M. (1978) Fm~tioMtion of nitrogen isotopes by animals A further complication to the use of variations in the natural abundance of N for tracer studies, J. Agric. Sci. 90, 7-9. STEWART W. D. P. ( 1978) Nitrogen-fixing cyanobacteria and their associations with eukaryotic plants. Endeavor 2, I70179. STUMP R. K. and FF~A~ER J. W. (1973) Simui~neous dete~ination of carbon, hydrogen and nitr~en in organic compounds. Nut. Sci. Abstr. 28, 746. SWARTP. K. ( 1983) Carbon and oxygen isotope fractionation in sclemctinian corals: a review. Earth-Science Reviews 19, 51-80. SWEENEY R. E., LIU K. K. and KAPLANI. R. (1978) Oceanic nitrogen isotopes and their uses in determining the source of sedimentary nitrogen. In Stable ~soto~s in the Earth Sciences (ed. 6. W. ROBINSON) pp. 9-26. DSIR. TAUBER H. Ii98 1) -Cevidence for dietarvhabits of nrebistoric man in Denmark. Nature 292, 332-j33. . THRESHERR. E. (1980) Reef Fish. Palmetto Pub. Co. VANDERMERWEN. J. (1982) Carbon isotopes, photosynthesis and archaeology. Amer. Scientist 70, 596-606. VAUGHNT. A. (1972) Mammo~o~. W. B. Saunders Co. VEIS A. f 1964) The Ma~romo~~lar Chemistry of Geletin. Academic Press. VIRGINIAR. A. and DELWICHEC. C. (1982) Natural N abundance of presumed N1-fixing and non-N2-fixing plants from selected ecosystems. Oecologia 54, 3 17-325. VOGELJ. C. (1978) Isotopic assessment of the dietary habits of ungulates. South Afr. J. Sci. 74, 298-307. WADA E. (1980) Nitrogen isotope fm~onation and its signilicance in biogeocfiemical processes occurring in marine environments. Isotope Marine Chemistry t&is. E. D. GOLDBERG, Y. HORIBEand K. SARUHASHI) pp. 375-398. Uchida Rokakuho Pubi. Co. Ltd. WADAE. and HA~ORI A. (1976) Natural abundance of i5N in particulate organic matter in the North Pacific Ocean. Geochim. Cosmochim. Acta 12.97-102. WADA E., KAD~NAGA T. and MATSUOS. (1975) N abundance in nitrogen of naturally muning substances and global assessment of denitrification from isotopic viewpoint. Geochem. J. 9, 139-148. WALKERE. P. (1975) Mammals ofthe World. Johns Hopkins University Press.