Journal of Archaeological Science (1995) 22, 327–340
Trophic Structure and Climatic Information From Isotopic
Signatures in Pleistocene Cave Fauna of Southern England
Hervé Bocherens* and Marilyn L. Fogel
Geophysical Laboratory, Carnegie Institution of Washington, 5251 Broad Branch Road N.W.,
Washington DC 200015-1305, U.S.A.
Noreen Tuross
Smithsonian Institution (CAL/MSC), 4210 Silver Hill Rd, Suitland, MD 20746, U.S.A.
Melinda Zeder
Smithsonian Institution, Anthropology, National Museum of Natural History, MRC 112, Washington DC 20560,
U.S.A.
(Received and accepted 4 September 1994)
The preservation of trophic structure and climatic information in Kent’s Cavern Upper Pleistocene mammal bones and
teeth was assessed by comparing the isotopic composition of modern and fossil equivalents. Yields of collagen from
both bone (N=19) and tooth (N=49) were extremely variable, with values relative to modern bone ranging from 0% to
100%. No evidence of preferential preservation of tooth collagen was detected. The carbon and nitrogen isotopic
diVerences in the herbivore versus carnivore collagen from Kent’s Cavern fauna were consistent with those observed in
modern faunas. Moreover, an enrichment of 0·4–1·7‰ in 15N was observed in tooth collagen of both deer and hyena
as compared to bone collagen. This enrichment presumably reflects a trophic shift from the consumption of milk during
infancy. Herbivores and carnivores had distinct diVerences in the carbon isotopic composition of enamel carbonate
hydroxylapatite, similar to those measured in modern specimens from similar climatic environments. The spacing
between the Ä13C values (diVerence between isotopic composition in collagen and carbonate hydroxylapatite) of Kent’s
Cavern herbivores and carnivores is similar to that measured in modern mammals from a single locality. The
preservation of primary oxygen isotopic composition of enamel carbonate hydroxylapatite was more diYcult to assess,
however. Oxygen isotopic compositions of Kent’s Cavern enamel are systematically lower than those of contempor-
aneous faunas from Southern France, which is consistent with a latitudinal eVect on rainfall oxygen isotopic
compositions. Although the Kent’s Cavern specimens have been subjected to extensive diagenetic alteration, the
biological isotopic signals seem to have utility for paleoecological reconstructions.
Keywords: STABLE ISOTOPES, CARBON, NITROGEN, OXYGEN, TOOTH, BONE, COLLAGEN, APATITE,
TROPHIC LEVEL, ISOTOPIC DIFFERENCES, KENT’S CAVERN, UPPER PLEISTOCENE.
Introduction
K ent’s Cavern is a Pleistocene cave formed inDevonian limestone located near the town ofTorquay, on the southern coast of Devonshire
(Great Britain). The largest excavation of this cave
took place from 1865 to 1880 under the direction of
William Pengelly. The tens of thousands of objects that
were removed during those 15 years have been dis-
persed to many museums world-wide, and the material
analysed in this work came from the zooarchaeological
collections of the Smithsonian Institution’s National
Museum of Natural History.
The samples discussed in this paper come from the
stratigraphic layer called ‘‘cave-earth’’ that yielded
an abundant fauna associated with Middle
Paleolithic (Mousterian) and Early Upper Paleolithic
(Aurignacian) artifacts (Garrod, 1926; Campbell,
1977). Four radiocarbon dates from the bone at the
‘‘cave-earth’’ layer (GrN6201, GrN6202, GrN6324 and
GrN6325) span 28,000 to 38,000 years bp (Campbell,
1977) and are in agreement with a more recent date
of 39,630&1420 bp (OxA-3403) obtained from
*Present address: Laboratoire de Biogéochimie Isotopique, Case
Courrier 120, Université P. et M. Curie, 4 Place Jussieu, F-75252
Paris Cedex 05, France.
327
0305–4403/95/020327+14 $08.00/0 ? 1995 Academic Press Limited
rhinoceros bone collagen. Spatial localization of the
bone samples within 1 m2 in the cave is possible due to
the detailed numbering on each specimen, and the
extensive records left by Pengelly. The majority of bone
samples analysed for this work were excavated in the
‘‘Smerdon’s Passage’’, an undervault close to the south
entrance, linking the ‘‘Passage of Urns’’ to ‘‘North
Sally Port’’. A few samples come from two other
places also located near the entrances of the cave in
the ‘‘cave-earth’’ layer, ‘‘North Sally Port’’ and
‘‘Vestibule’’.
The ‘‘cave-earth’’ layer contains a very abundant
Devonsian fauna of large mammals (Garrod, 1926;
Campbell, 1977). This fauna includes carnivores, such
as hyena (Crocuta crocuta) in great abundance, and
herbivores, such as woolly rhinoceros (Coelodonta
antiquitatis), equid (Equus sp.), bison (Bison sp.) and
wild ox (Bos primigenius), reindeer (Rangifer tarandus),
red deer (Cervus elaphus) and giant elk (Megaloceros
giganteus), mammoth (Mammuthus primigenius), as
well as brown bear (Ursus arctos: Campbell (1977) has
deleted the cave bear Ursus spelaeus from the last
glacial faunal list because this species occurs in situ
only in the earlier ‘‘breccia’’ layer). The samples analy-
sed in this study are presented in Tables 1 & 2. The diet
and paleoecology of these species are already well
known from their morphology, in some cases frozen
stomach contents. Some of these species are living
today, often in diVerent climatic conditions.
Based on stomach and intestine contents, it is
known that horse and woolly rhinoceros were eating
mostly grasses in the Eurasian arctic (Vereshchagin &
Baryshnikov, 1982). The summer diet of mammoths,
inferred from fossilized stomach contents found in
frozen carcasses, was composed of herbaceous plants,
mosses, bushes and tree twigs and bark (Vereshchagin
& Baryshnikov, 1982, 1984; SutcliVe, 1985). The red
deer is very adaptable and exhibits a considerable
variation in size. Further, the red deer was capable of
living in forests as well as in open country (Chaplin,
1975; Guthrie, 1982). The giant deer was most prob-
ably an opportunistic browser that supplemented
its diet with large amounts of grass (Barnosky,
1986). Wild ox lived in open country and open forest
(Anderson, 1984). The spotted hyena was the top
carnivore and scavenger of this ecosytem, by compari-
son with the living ones in Africa (Vereshchagin
& Baryshnikov, 1984). Finally, brown bears were
omnivorous and their diet may be mostly composed
of plant material or meat according to the availability
of these sources of food (Dendaletche, 1982).
Trophic structure and climatic information can be
recorded in the isotopic composition of vertebrate
bones and teeth, in bone and dentin organic matter
(13C, 15N) and in enamel carbonate hydroxylapatite
(13C, 18O). If preserved, this isotopic signature provides
a way to retrieve the structure of ancient trophic webs
and to reconstruct paleoclimates. One test of preser-
vation of the isotopic signals is the comparison of some
specific isotopic signatures in fossil samples with those
in modern equivalents. The diet of a mammal is
reflected in the isotopic composition of animal tissues
based on the trophic level of the individual as an adult,
and on the shift of diet from mother’s milk to adult diet
at weaning. The trophic level eVect leads to a diVerence
in bone collagen carbon and nitrogen isotopic abun-
dances (Schoeninger & DeNiro, 1984; van der Merwe,
1989) and in enamel carbonate hydroxylapatite car-
bonate isotopic abundances between an animal and its
diet (Krueger & Sullivan, 1984; Lee-Thorp et al., 1989).
The shift of diet at weaning (i.e. the nursing eVect)
leads to a diVerence in nitrogen isotopic abundances
(Fogel et al., 1989, Tuross & Fogel, 1994) which is
recorded as a diVerence between bone and dentin
collagen of an individual in species where teeth stop
their growth shortly after weaning time (Bocherens
et al., 1992, 1994). Individuals from species where teeth
continue to grow after weaning contain 15N abun-
dances similar to those in bone, because collagen is
synthesized from the adult diet in both teeth and bones
(Bocherens et al., 1992, 1994).
Climatic conditions also influence the isotopic sig-
nals in mammal mineralized tissues. The nitrogen
isotopic composition of collagen can be influenced by
water and food availability through a trophic system
(Heaton et al., 1986; Sealy et al., 1987; Ambrose, 1991).
Carbon isotopic composition of carbonate hydroxyl-
apatite of herbivore enamel has been shown to depend
on the climate: for similar collagen carbon isotopic
compositions, the enamel carbon isotopic abundances
in herbivores from cold areas are more enriched in 13C
relative to those of herbivores from South Africa
(Bocherens & Mariotti, 1992). Finally, the oxygen
isotopic composition of carbonate hydroxylapatite is
linked to those of the drinking water, and thus to the
temperature (Koch et al., 1989).
Kent’s Cavern ‘‘cave-earth’’ deposit provides a
favorable setting to investigate the preservation of the
isotopic signal of trophic structure in bone organic
matter (13C, 15N) and enamel (13C, 18O), by comparing
the isotopic values between herbivorous and carnivor-
ous animals, isotopic values in bone and teeth of
jawbones from deer and hyenas, and by comparing the
isotopic values obtained in Kent’s Cavern samples with
those measured in French Upper Pleistocene localities
already published (Bocherens et al., 1991a,b, 1994;
Fizet et al., 1994). We tested the preservation of the
isotopic composition in the organic matter extracted
from bone and dentin, and in the carbonate hydroxyl-
apatite of enamel. The carbonate hydroxylapatite of
bone and dentin has been shown to be isotopically
altered in the early diagenesis (Lee-Thorp, 1989; Koch
et al., 1990) and thus will not be considered here. In
summary, the diverse fossil mammal fauna of the
Upper Pleistocene of Kent’s Cavern was used to test if
the isotopic compositions are preserved by comparing
the results with those obtained on modern and on fossil
equivalents.
328 H. Bocherens et al.
Table 1. Isotopic composition of collagen and of enamel carbonate hydroxylapatite of herbivores from Kent’s Cavern
Sample
no. 1
Sample
no. 2 Layer Taxon Sample
Yield
(mg g"1)
Collagen Apatite
C/N
ä13C
(‰)
ä15N
(‰)
ä13C
(‰)
ä18O
(‰)
Ää13C
(‰)
214264 n.d. Equus M3 112·2 3·4 "20·8 4·4
61351 12/3468 SP Equus tooth 172·8 3·5 "21·2 7·8 "10·2 26·3 11·0
61370 10/4849 NSP Equus upper tooth 92·4 3·9 "22·0 2·2 "11·6 26·9 10·4
61372 4/4871 NSP Equus metacarpal 2·7 3·5 "21·2 8·4
61375 9/5308 SP Equus tooth 9·3 n.d. "21·6 6·8 "10·9 24·2 10·7
61376 SP Equus tooth 81·5 3·5 "21·6 7·7
61376 6/5315 SP Equus tooth n.d. 3·7 "20·4 7·2 "11·5 24·9 9·9
61376 7/5315 SP Equus tooth 6·8 4·0 "21·4 4·6
61376 8/5315 SP Equus tooth "12·0 24·8
61377 16/3389 SP Equus upper tooth 94·2 3·4 "21·9 6·9
61378 9/3400 SP Equus tooth 107·2 3·5 "21·6 6·0 "10·5 25·9 11·1
61381 2/3461 SP Equus tooth 4·5 3·5 "21·4 3·3
61386 8/3528 SP Equus lower tooth 171·5 3·6 "21·0 8·5 "11·5 26·7 9·5
61387 2/3536 SP Equus tooth 1 "11·3 26·4
4/3536 SP Equus tooth 158·8 3·4 "21·6 5·4 "11·4 25·0 9·7
Average "21·4 6·5 "11·2 25·7 10·3
s.d. &0·3 &1·7 &0·6 &1·0 &0·6
214255 ac38778-1 n.d. Coelodonta long bone 183·7 3·4 "20·6 3·7
61369 2/4809 NSP Coelodonta scapula 16·5 no N b.p. b.p.
4860 NSP Coelodonta bone 4·3 3·5 "20·6 5·9
61372 4871 NSP Coelodonta bone 32·9 no N b.p. b.p.
61374 1/5297 SP Coelodonta astragalus 47·3 3·3 "20·6 4·4
61377 1/3389 SP Coelodonta metapodial 33·4 3·8 "21·2 6·6
61381 3461 SP Coelodonta bone 91·2 3·5 "22·5 6·7
61389 1/3643 SP Coelodonta bone 0
61389 3643 SP Coelodonta bone 118·2 5·7 b.p. b.p.
Average (bones) "21·1 5·2
s.d. &0·9 &1·4
61369 NSP Coelodonta tooth 168·1 3·3 "20·2 8·3 "8·9 26·0 11·3
61376 24/5315 SP Coelodonta tooth 142·1 3·5 "20·7 6·4
61377 11/3389 SP Coelodonta* cement 161·8 3·4 "20·8 4·7 "11·0 26·0 9·8
61377 11/3389 SP Coelodonta* dentin 151·1 3·4 "20·2 4·3 "11·0 26·0
61378 21/3400 SP Coelodonta tooth 169·4 3·3 "20·7 6·0
61381 6/3461 SP Coelodonta tooth 164·3 3·4 "21·0 4·4 "10·8 23·2 10·2
4/3470 SP Coelodonta* dentin 179·8 3·5 "20·2 8·0 "10·2 24·0 10·0
4/3470 SP Coelodonta* cement 52·5 n.d. "20·7 8·2 "10·2 24·0 10·5
61387 5/3536 SP Coelodonta tooth n.d. 3·4 "20·2 5·7 "10·7 26·8 9·5
60/2104 V Coelodonta tooth 25·7 n.d. "21·0 9·1 "12·0 27·1 9·0
Average (teeth) "20·6 6·5 "10·6 25·5 10·0
s.d. &0·3 &1·8 &1·0 &1·6 &0·7
214257 n.d. cervid* maxillary bone 15·1 n.d. "21·0 5·1
61371 15/4860 NSP cervid radius 61·8 3·5 "23·3 5·3
61378 5/3400 SP cervid astragalus 16·2 3·6 "23·4 5·5
Average (bones) "22·6 5·3
s.d. &1·4 &0·2
214257 n.d. cervid* M2 103·7 3·4 "20·1 6·7 "11·1 25·7 9·0
214257 n.d. cervid* M3 65·1 3·4 "20·5 6·4 "11·1 26·7 9·4
61369 4/4809 NSP cervid tooth 8·9 "9·9 21·5
61370 13/4849 NSP cervid* M2 44·4 3·7 "21·0 8·0 "11·1 26·6 9·9
61370 13/4849 NSP cervid* M3 68·6 3·5 "20·3 7·4 "10·9 24·9 9·4
61378 11/3400 SP cervid tooth 166·6 3·5 "20·6 5·0
61378 12/3400 SP cervid tooth 115·6 3·4 "21·1 7·3
Average (teeth) "20·5 6·6 "10·8 25·1 9·4
s.d. &0·4 &1·1 &0·5 &2·1 &0·4
214257 n.d. Bos M3 115·6 3·5 "20·6 5·6
6138? n.d. Mammuthus molar "10·3 26·7
SP, Smerdon’s Passage; NSP, North Sally Port; V, vestibule; sample no. 1, Smithsonian’s numbers; sample no. 2, Pengelly’s numbers.
*, Samples from the same individual.
n.d., ‘‘Not determined’’; b.p., ‘‘Bad preservation’’, in cases where samples had a clayish aspect after extraction and yielded very little quantities
of gases.
Ä13C=ä13Capatite" ä
13Ccollagen.
Isotopic Signatures From Pleistocene Cave Fauna 329
Methods of Analysis
To obtain the insoluble organic matter, chunks of bone
were demineralized with 0·5 m EDTA, pH 7·2, at 4)C.
The residues were washed 15 times with distilled water
and lyophilized. When the organic matter is fairly well
preserved, the result is a translucent, pale yellow col-
lagen replica (Tuross et al., 1988). Isotopic measure-
ments were performed on gases extracted and purified
after a ‘‘Dumas’’ combustion as described in Estep &
Vigg (1985). One to three milligrams of collagen were
combusted with copper oxide and metal copper in a
sealed quartz tube at 900)C for 1 h. The combustion
tube was cooled at a defined rate, and the products of
combustion were isolated by cryogenic distillation. On
most of the samples, the determination of the C/N
atomic ratio was determined with a Carlo-Erba
EA1108 elemental analyzer. The range of C/N values
for unaltered samples is presumed to be 2·9–3·6
(DeNiro 1985).
The enamel samples were pretreated according to
Bocherens et al. (1991b). The powdered enamel was
treated by NaOCl 2–3% during 20 h at 20)C, rinsed
carefully with distilled water and then treated with a
1 m buVered solution of acetic acid (pH 4·75) for 20 h
at 20)C to remove any exogenous carbonate without
dissolving more than 10% of the enamel sample. The
CO2 was then extracted from the inorganic powders by
orthophosphoric acid at 50)C for 5 h, following Koch
et al. (1989).
The isotope ratios are expressed for carbon as ä13C
versus PDB (a marine carbonate), for nitrogen as ä15N
versus atmospheric N2 and for oxygen as ä
18O verus
SMOW (standard mean oceanic water):
äEX=
Rsample "Rstandard#1000,
Rstandard
where äeX=ä13C and 13R=13C/12C; äEX=ä15N and
15R=15N/14N and äEX=ä18O and 18R=18O/16O res-
pectively. Carbon and oxygen isotopic ratios were
measured on a Finnigan MAT 252, and nitrogen
isotopic ratios were measured on a Nier-Johnson type
double focusing mass spectrometer. Analytical pre-
cisions were 0·1‰ for ä13C, 0·2‰ for ä15N and 0·2‰
for ä18O values. The correction is not known for
oxygen in carbonate hydroxylapatite, therefore the
correction formula for calcite at 50)C was used (Koch
et al., 1989).
Results
Collagen yields in bone and dentin
The yields of the organic matter, defined as the ratio of
dry weight of organic matter after decalcification to the
Table 2. Isotopic composition of collagen and enamel carbonate hydroxylapatite of carnivores and omnivores from Kent’s Cavern
Sample
no. 1
Sample
no. 2 Layer Taxon Sample Yield (mg g"1)
Collagen Apatite
C/N
ä13C
(‰)
ä15N
(‰)
ä13C
(‰)
ä18O
(‰)
Ää13C
(‰)
ac38778-4 n.d. Crocuta* maxillary bone 159·0 3·3 "18·9 8·0
61375 15/5308 SP Crocuta* mandible 58·3 3·3 "20·3 9·3
9/5315 SP Crocuta* mandible 45·7 3·5 "20·3 11·2
61376 12/5315 SP Crocuta* maxillary bone 205·3 3·3 "19·4 10·8
61384 9/3478 SP Crocuta* mandible 25·6 3·7 "20·0 9·0
61387 8/3536 SP Crocuta* mandible 145·6 3·4 "19·0 9·0
Average (bones) "19·6 9·7
s.d. &0·7 &1·3
ac38778-4 n.d. Crocuta* P3 198·3 3·3 "18·8 8·4 "12·8 25·8 6·0
214259 n.d. Crocuta M1 47·3 3·6 "19·7 13·0 "13·4 25·6 6·3
214260 20/3446 SP Crocuta upper C 62·0 3·4 "19·6 9·4
61375 15/5308 SP Crocuta* P3 170·8 3·4 "20·4 11·0
9/5315 SP Crocuta* P4 50·2 3·4 "19·1 11·7 "12·8 26·0 6·3
9/5315 SP Crocuta* P3 95·2 3·4 "18·9 12·2 "13·5 25·9 5·4
61376 12/5315 SP Crocuta* P4 130·9 3·2 "19·3 11·9 "13·5 25·2 5·8
61377 6/3389 SP Crocuta P3 25·0 3·5 "19·1 9·5
61382 2/3470 SP Crocuta dentin 0
61384 9/3478 SP Crocuta* M1 112·4 3·4 "19·9 9·4 "13·6 25·2 6·3
61384 9/3478 SP Crocuta* P4 71·4 3·5 "19·9 9·5 "14·1 25·3 5·8
61384 10/3478 SP Crocuta canine 86·4 3·6 "19·8 9·3 "13·5 25·2 6·2
61386 3/3528 SP Crocuta canine 36·6 3·5 "19·4 11·7
61386 4/3528 SP Crocuta dentin 123·0 3·4 "19·3 9·5
61387 8/3536 SP Crocuta* M1 95·9 3·3 "19·2 9·7 "12·5 22·1 6·7
Average (teeth) "19·5 10·4 "13·3 25·1 6·1
s.d. &0·4 &1·4 &0·5 &1·2 &0·4
61384 3/3478 SP Ursus arctos canine 140·1 3·6 "20·0 12·6 "13·9 26·3 6·1
Key as for Table 1.
330 H. Bocherens et al.
dry weight of fossil bone or tooth, were highly variable
within the deposit. A majority of samples yielded a
very well preserved collagen, in the form of a yellowish
replica of the original sample, with C/N values between
3·2 and 3·6. Only two samples out of 66 yielded no
residue after EDTA dissolution. In three samples, the
final product consisted of only insoluble minerals,
probably clay, that generated only minute quantities of
gases after combustion. The C/N values were clearly
outside the biological range (no nitrogen for samples
nos. 61369-2/4849 and 61372-4871, C/N=5·7 for
sample no. 61389-3643; Table 1). In the well-preserved
samples, the amount of collagen was sometimes
very close to the amount in fresh bone (around
200 mg g"1); such samples have retained almost all
their original collagen. The yields in organic matter
extracted from bones and teeth are similar (Tables 1 &
2 and Figure 1), although a majority of bone samples
yielded less than 60 mg g"1 and the bulk of dentin
samples contained more than 80 mg g"1 (Figure 1). A
comparison of the yields of bone and dentin from the
same jawbone specimens, however, indicates that bone
did not always contain less collagen than dentin (see
Table 4). Thus in terms of total collagen content, bone
was preserved as well as dentin.
The range of yield in collagen demonstrates the
variability of organics within sites, which has been
described in other localities (Bocherens et al., 1991a,b;
Tuross & Stathoplos, 1993). Only a few samples with
‘‘biological isotopic values’’ (see discussion below) had
elevated C/N values (3·7–4·0). Most of the samples
with C/N values greater than 3·6 had collagen yields
lower than 50 mg g"1, with one exception (horse, no
61370-10/4849: yield of 92·4 mg g"1 and C/N=3·9;
Table 1). Samples with collagen yields lower than
10 mg g"1 typically had C/N values lower than 3·6.
The isotopic compositions measured on the samples
with C/N greater than 3·6 are in italics in the tables.
Isotopic compositions of these compromised samples
were not used for the calculation of average values.
Isotopic compositions in collagen
In herbivores, the ä13C values of collagen (C/N values
3·2–3·6) range from "23·4 to "20·1‰ with a mean of
ä13C="21·0&0·9 (n=33) (Table 1 and Figure 2). In
carnivores, the ä13C values of collagen range from
"20·4 to "18·8‰ with a mean of "19·5&0·5‰
(n=20) in carnivores (Table 2 and Figure 2). The ä13C
values between bone and tooth collagen for a given
species are not significantly diVerent (Student t-test,
P>0·1). In the deer and hyena jawbones, ä13C values of
bone and tooth collagen within an individual were
virtually indistinguishable and the diVerence was al-
ways less than 1‰ (Tables 3 & 4, Figure 3). A paired
t-test performed on the bone and tooth collagen ä13C
values of the hyena jawbones indicated that these
values were not significantly diVerent (P>0·1). Thus,
the within-species ä13C values of bone and tooth
collagen will be considered together. On average, there
is an increase of &1·5‰ between the ä13C values of the
carnivores relative to the ä13C values of the herbivores
among Kent’s Cavern samples. This diVerence is
statistically significant (Student’s t-test, P<0·01). Two
of the ä13C values measured on deer (Table 1; nos.
61371-15/4860 and 61378-5/3400) are depleted relative
to the majority of herbivore bone collagen.
Herbivores ä15N values in bone and tooth collagen
(C/N values 3·2–3·6) ranged from 3·3 to 8·4‰
(Figure 2) and averaged ä15N=6·3&1·6‰ (N=33).
ä15N values ranged in carnivores from 8·0 to 13·0‰
(Figure 2), and averaged ä15N=10·2&1·4‰ (N=20).
The diVerence of 3·9‰ between collagen average ä15N
value in herbivores and carnivores is statistically sig-
nificant (Student’s t-test, P<0·01). In rhinoceros, deer
and hyenas, species with a diVerent number of samples
of teeth and bones for an intra-individual comparison,
the ä15N values in bone collagen were on average lower
than those of tooth collagen (Tables 1 & 2; 5·2&1·4‰
versus 6·5&1·8‰ in the rhinoceros; 5·3&0·2‰ versus
6·6&1·1‰ for deer; 9·7&1·3‰ versus 10·4&1·4‰ for
hyenas). Given the large range of ä15N values, there
was no statistically significance in these diVerences.
The single ä15N value of horse bone collagen was
within the range of ä15N values of horse tooth collagen
(Table 1; 8·4‰ versus 3·3–8·5‰). In deer and hyenas,
the comparison of ä15N values in bone and tooth
collagen from the same jawbones indicates that tooth
collagen is always more positive than bone collagen
from the same individual, from 0·4 to 1·7‰ (Table 3
and Figure 3). Moreover, a paired t-test performed on
the bone and tooth collagen ä15N values of the hyena
jawbones showed that these values were significantly
diVerent (P<0·01).
The ä13C and ä15N values of the degraded collagen
with C/N values greater than 3·6 were in most of the
cases within the range of the isotopic compositions
measured on samples with C/N values between 3·2
and 3·6. A horse tooth (sample no. 61372-4/4871)
was the exception; ä13C and ä15N were lower than all
0
Collagen yield (mg g–1)
N
u
m
be
r 
of
 s
am
pl
es
0–
20
8
4
2
10
6
20
–4
0
20
0–
22
0
18
0–
20
0
16
0–
18
0
14
0–
16
0
12
0–
14
0
10
0–
12
0
80
–1
00
60
–8
0
40
–6
0
Bone
Dentin
Modern bones
and teeth
Figure 1. Collagen yields in bone and dentin of Kent’s Cavern
mammals.
Isotopic Signatures From Pleistocene Cave Fauna 331
d13C (‰)
–18–24 –22 –20–23 –21 –19
Carnivore teeth
Carnivore bones
Herbivore teeth
Herbivore bones + horse teeth
d
15N (‰)
0
Carnivore teeth
Carnivore bones
Herbivore teeth
Herbivore bones + 
horse teeth
1413121110987654321
Rhinoceros bone
Rhinoceros tooth
Equid bone + tooth
Cervid bone
Bovid tooth
Hyena bone
Hyena tooth
Cervid tooth
Tr
op
h
ic
 le
ve
l
Figure 2. Variations of ä13C and ä15N values in Kent’s Cavern mammal tooth and bone collagen according to the trophic level as determined
from anatomical data. Only samples with C/N values comprised between 3·2 and 3·6 have been plotted.
Table 3. Isotopic composition of collagen from bone and tooth of individual deer from Kent’s Cavern
Sample
no. 1
Sample
no. 2 Layer Taxon Sample
Yield
(mg g"1)
Collagen
Ää13C
(‰)
Ää15N
(‰)C/N
ä13C
(‰)
ä15N
(‰)
214257 n.d. cervid max. bone 15·1 n.d. "21·0 5·1
214257 n.d. cervid M2 103·7 3·4 "20·1 6·7 0·9 1·6
214257 n.d. cervid M3 65·1 3·4 "20·5 6·4 0·5 1·3
61370 13/4849 NSP cervid M2 44·4 3·7 "21·0 8·0
61370 13/4848 NSP cervid M3 68·6 3·5 "20·3 7·4
Key as for Table 1.
Ää13C=ä13Ctooth "ä
13Cbone; Ää
15N=ä15Ntooth "ä
15Nbone.
332 H. Bocherens et al.
other horse isotopic compositions (Table 1, and see
Figure 5).
Isotopic compositions in carbonate hydroxylapatite
The ä13C values of herbivore enamel from Kent’s
Cavern ranged from "12·0 to "8·9‰ ("10·9&
0·7‰, N=20; Table 1 and Figure 4). In contrast, the
one carnivore group, hyenas, had ä13C values from
enamel carbonate hydroxylapatite in range of "14·1
to "12·5‰ ("13·3&0·5‰, N=9; Table 2 and Figure
4). This range did overlap with herbivores. The diVer-
ence of the average carbonate hydroxylapatite ä13C
values between herbivores and carnivores was statisti-
cally significant (Student’s t-test, P<0·01). The ä13C
values measured on rock matrix and recrystallized
calcite (Table 5 and Figure 4) were more positive than
those measured on the enamel samples, ranging from
"8·7 to "7·5‰.
The diVerence in the ä13C values between collagen
and carbonate hyroxylapatite (Ä13C) ranged from 9·0
to 10·7‰ in Kent’s Cavern herbivores (Table 1). The
Table 4. Isotopic composition of collagen from bone and tooth of the same hyena individuals from Kent’s Cavern
Sample
no. 1
Sample
no. 2 Layer Taxon Sample
Yield
(mg g"1)
Collagen
Ää13C
(‰)
Ää15N
(‰)C/N
ä13C
(‰)
ä15N
(‰)
ac38778-4 n.d. Crocuta max. bone 159·0 3·3 "18·9 8·0
ac38778-4 n.d. Crocuta P3 198·3 3·3 "18·8 8·4 0·1 0·4
61375 15/5308 SP Crocuta mandible 58·3 3·3 "20·3 9·3
61375 15/5308 SP Crocuta P3 170·8 3·4 "20·4 11·0 "0·1 1·7
9/5315 SP Crocuta mandible 71·6 3·5 "19·6 11·1
9/5315 SP Crocuta P4 50·2 3·4 "19·1 11·7 0·5 0·6
9/5315 SP Crocuta P3 95·2 3·4 "18·9 12·2 0·3 1·1
61376 12/5315 SP Crocuta max. bone 205·3 3·3 "19·4 10·8
61376 12/5315 SP Crocuta P4 130·9 3·2 "19·3 11·9 0·1 1·1
61384 9/3478 SP Crocuta mandible 25·6 3·7 "20·0 9·0
61384 9/3478 SP Crocuta M1 112·4 3·4 "19·9 9·4 0·1 0·4
61384 9/3478 SP Crocuta P4 71·4 3·5 "19·9 9·5 0·1 0·5
61387 8/3536 SP Crocuta mandible 145·6 3·4 "19·0 9·0
61387 8/3536 SP Crocuta M1 95·9 3·3 "19·2 9·7 "0·2 0·7
Average 0·1 0·8
s.d. &0·2 &0·4
Key as for Table 1.
Ää13C=ä13Ctooth "ä
13Cbone; Ää
15N=ä15Ntooth "ä
15Nbone.
Recent bear
+13–22
Hyena 15/5308
Reindeer (Lapland)
Hyena 38778–4
Hyena 9/5315
Hyena 8/3536
Recent wolf
Deer (Kent's Cavern)
Reindeer (Alaska)
Hyena 12/5315
–20 +1 +3 +5–21 –19 +2 +4 +12
Hyena 9/3478
d
15N (‰)
+6 +7 +8 +9 +10 +11
d
13C (‰)
Figure 3. Variations of ä13C and ä15N values in Kent’s Cavern mammal tooth and bone collagen of deer and hyena jawbones compared to
those in modern equivalents (values for modern reindeers and carnivores from Bocherens, 1992). The point for the hyena no. 9/3478 is in grey
because of its high C/N value (3·7).
Isotopic Signatures From Pleistocene Cave Fauna 333
average Ä13C values were 10·3&0·6‰ (N=7) in
horses, 10·0&0·7‰ (N=7) in woolly rhinoceros,
9·7&0·3‰ (N=4) in deer. There was no significant
diVerence in the Ä13C values between these diVerent
species of herbivores (Student’s t-test, P<0·01). In
carnivores, Ä13C values ranged from 5·8 to 6·7‰
(Table 2) and the average value was 6·1&0·4‰ (N=9).
The diVerence between the average Ä13C values
of carnivores and each species of herbivores was
statistically significant (Student’s t-test, P<0·01).
The oxygen isotopic values ranged from 21·5 to
27·1‰ (Tables 1 & 2 and Figure 4). The average values
of ä18O were not significantly diVerent between horses
(ä18O=25·7&1·0‰), woolly rhinoceros (ä18O=25·5&
1·6‰), deer (ä18O=25·1&2·1‰) and hyenas (ä18O=
25·1&1·2‰). The ä18O values measured on samples of
rock matrix surrounding the fossils and on some
diagenetic calcite formed within a horse tooth were
slightly higher than those found in the fossil specimens
(27·2–27·9‰; Table 5 and Figure 4).
Discussion
Preservation of the carbon and nitrogen isotopic
compositions in organic matter
In addition to the C/N ratio of the extracted organic
matter (DeNiro, 1985), one test for the preservation of
the isotopic composition of ancient organic matter is
to compare those measured to equivalent modern
material. As stated previously, two kinds of isotopic
diVerences in the bone and tooth collagen have been
30
20
–7–15
24
22
28
26
–11 –9–12 –10 –8
d
13C (‰)
d
18
O
 (
‰
)
Bear
Hyena
Mammoth
Deer
Rhinoceros
–13–14
x
Equid
Matrix+
Figure 4. ä13C and ä18O values in Kent’s Cavern mammal enamel carbonate hydroxylapatite compared to those of rock matrix.
Table 5. Isotopic composition of carbonate from rock matrix surrounding Kent’s Cavern samples
Sample
no. 1
Sample
no. 2 Location Sample
ä13C
(‰)
ä18O
(‰)
61384 SP matrix "8·7 27·2
61387 8/3536 SP matrix around hyena tooth "8·4 27·4
4/3536 SP calcite in equid tooth "7·5 27·9
Average "8·2 27·5
s.d. &0·6 &0·4
Key as for Table 1.
334 H. Bocherens et al.
identified in mammals: (1) the diVerence between
herbivores and carnivores; and (2) the diVerence
between bone and dentin collagen of the same indi-
viduals in the species where teeth stop their growth
around weaning time. A clear enrichment in 15N has
been shown to exist between herbivore and carnivore
collagen. This enrichment is reported to range in
terrestrial ecosystems from 2·8‰ (Schwarcz, 1991) to
5·7‰ (Ambrose & DeNiro, 1986) with an average
value between 3 and 4‰ (Schoeninger & DeNiro,
1984; Schoeninger, 1985; Sealy et al., 1987). A less
obvious enrichment in organic 13C, up to 2‰, has
also been reported (van der Merwe, 1989; Lee-Thorp
et al., 1989), but has not been observed in all terres-
trial ecosystems studied to date. The diVerence in
organic ä13C values between herbivores and carni-
vores seems to be observed in ecosystems where all
the plants have a C3 photosynthetic pathway, and
thus present a small range of carbon isotopic input
(van der Merwe, 1989). The other isotopic diVerence
in modern and fossil mammals is the enrichment in
15N in dentin collagen relative to bone collagen in
species where teeth stop growing around weaning
time (e.g. reindeer, bear, wolf). The ä15N of bone
and tooth overlaps in species with continuously-
growing teeth, such as horse (Bocherens et al., 1992,
1994).
In Kent’s Cavern samples, the isotopic compositions
(ä13C and ä15N) are within the range of values
measured in modern mammals from temperate and
cold areas (Tauber, 1986; Bada et al., 1990; Bocherens
et al., 1994). For instance, the average ä13C value of
modern European herbivores published in Tauber
(1986) range from "21·8 to "20·4‰ for equivalent
species than in Kent’s Cavern, and the average ä13C
value of modern European herbivores published in
Bocherens et al. (1994) is "21·1&0·9‰ (N=7) (range
from "22·1 to "19·9‰). Furthermore, isotopic com-
positions of Kent’s Cavern material are similar to those
measured on the same species or on species of the same
trophic level in other European Upper Pleistocene
localities (Bocherens et al., 1991a,b, 1994; Fizet et al.,
1994). Two of the ä13C values measured on deer from
Kent’s Cavern, however, are lower than the ä13C
values measured on the majority of herbivore bone
collagen ("23·3‰ and "23·4‰ for deer bone nos.
61371-15/4860 and 61378-5/3400, respectively; Table
1). The collagen yield on these two samples is com-
parable to others, the C/N values are in the
upper acceptable range of ratios (3·5 and 3·6) and
the ä15N values are typical of other herbivores. Such
negative ä13C values have been measured only on
deer samples coming from two diVerent places in the
cave, and not on other species. Therefore, we have no
basis for rejecting these values, although analogous
results in other deer samples and a careful examina-
tion of the biochemical composition of these collagen
samples is necessary to confirm the more negative
ä13C values.
In comparing the ä15N values, those measured in
bones and in dentin of species with teeth formed only
during infancy are considered first (Bocherens et al.,
1992, 1994). For this purpose, we compared the
values from the tooth and bone collagen of one deer
mandible (Table 3) and six hyena fragmentary jaw-
bones (Table 4). For these two taxa, higher values in
tooth relative to bone collagen were observed, as in
the closely related living groups (reindeer and carni-
vores such as wolf and black bear, respectively) with
a similar dietary pattern. The ä15N values were lower
in bone collagen than in dentin collagen in deer, as
observed in modern and fossil reindeer (Figure 2,
Bocherens et al., 1994) and in hyenas, as observed in
modern carnivores such as wolf and black bear (Fig-
ure 2; Bocherens et al., 1994), and in fossil cave bear
(Bocherens et al., 1994). The diVerences between the
average ä15N values of bone and tooth collagen for
deer and hyenas are very similar to those measured
on individual jaws (1·6‰ for deer and 0·8‰ for
hyenas). When two teeth from the same jawbone
were analysed (Tables 3 & 4; deer nos. 214257 and
13/4849, hyena nos. 9/5315 and 9/3478), the diVerence
of ä15N values are rather small (0·1–1·0 per thousand)
and of the same magnitude than for modern reindeers
and carnivores (Figure 2; Bocherens, 1992). Thus, for
these two species, a diet of milk produces higher ä15N
values in tooth than in bone collagen. The ä15N
values of these tissues are, therefore, considered
separately in the discussion of the trophic level of
these species.
Measurements performed on modern and fossil in-
dividuals demonstrated that the ä15N values of horse
dentin collagen reflect the adult diet in a fashion
analogous to bone collagen (Bocherens et al., 1992,
1994; Fizet et al., 1994). For the purpose of a trophic
level discussion, the ä15N values of Kent’s Cavern
horse tooth collagen were thus reported with the ä15N
values of bone collagen of the other herbivore species.
In the case of the woolly rhinoceros, the ä15N values
measured on these two tissues were separated because
of the type of tooth growth and the slightly higher
average ä15N value of the dentin collagen relative
to the average ä15N value of the bone collagen
(6·5&1·8‰ versus 5·2&1·4‰, respectively). Future
work will resolve the variability of the tissues with
more certainty.
We consider only the values measured on rhinoceros
bone collagen for the discussion of trophic level. It
appears that the diVerence between the average ä15N
value in herbivore bone (including horse tooth) and
carnivore bone (6·0&1·6 and 9·7&1·3, respectively) is
3·7‰, almost the same as the diVerence between the
average ä15N value in herbivore and carnivore teeth
(6·5&1·5 and 10·4&1·4, respectively). This diVerence
of 3·7‰ is very similar to others measured between
mammal herbivores and carnivores in most modern
terrestrial ecosystems, usually between 3 and 4‰
(Schoeninger & DeNiro, 1984; Schoeninger, 1985;
Isotopic Signatures From Pleistocene Cave Fauna 335
Sealy et al., 1987; Tuross et al., 1994). The ä15N
values in herbivores and carnivores in the total popu-
lation overlap slightly (Figure 1), but when both
carbon and nitrogen isotope ratios are considered,
herbivores and carnivores form two distinct groups
(Figure 5). A comparison of the mean&1 s.d. values
in Kent’s Cavern and three French Upper Pleistocene
localities (Figure 6) shows a similar pattern of vari-
ations, with lower ä13C values in herbivores and
carnivores, and higher ä15N values in carnivores than
in herbivores. We conclude that the similarity of the
isotopic compositions between fossil and modern
equivalents and other fossil material reflects the
preservation of some of the biogenic signal.
Preservation of isotopic compositions in carbonate
hydroxylapatite
The ä13C values of carbonate generated from hydroxy-
lapatite of herbivores from Kent’s Cavern ("12·0 to
"8·9‰; Table 1 and Figure 4) overlap the range
reported by Bocherens & Mariotti (1992) for European
and Alaskan modern large herbivores ("13·7 to
"10·3‰) and for European and Alaskan Pleistocene
large herbivores ("12·4 to "10·1‰). Only one value
for a rhinoceros sample ("8·9‰ for sample no. 61389)
is slightly more positive. The values measured on rock
matrix and recrystallized calcite (Table 5 and Figure 4)
are less negative than those measured on the enamel
samples, with no overlap ("8·7 to "7·5‰). In hyenas
from Kent’s Cavern, the ä13C values of enamel
carbonate hydroxylapatite ("13·3&0·5‰, range
from "14·1 to "12·5‰) are comparable to those
measured in Pleistocene carnivores from other
localities (Bocherens et al., 1994). In the fossil samples
from French localities, the ä13C values of enamel
carbonate hydroxylapatite ranged from "12·6 to
"12·1‰ for carnivores (Bocherens, 1992; Bocherens
et al., 1994). However, they are slightly less negative
than for modern carnivores from South Africa
with similar collagen ä13C values ("16 to "13‰;
Lee-Thorp et al., 1989) and for the very few specimens
of carnivores from temperate areas analysed to date
("15·3‰; Bocherens & Mariotti, 1992). Nevertheless,
the diVerence between the ä13C values of enamel
carbonate hydroxylapatite between herbivores and
carnivores (2·6‰) is very similar to those measured
in modern mammals from defined localities (Bocherens
14
0
–18–24
5
8
6
4
2
1
9
7
3
–22 –20–23 –21 –19
d
13C (‰)
10
11
12
13
d
15
N
 (
‰
)
Herbivore bone + horse tooth (3·2≤C/N≤3·6)
Bear tooth (3·2≤C/N≤3·6)
Hyena tooth (3·2≤C/N≤3·6)
Hyena bone (3·6<C/N)
Hyena bone (3·2≤C/N≤3·6)
Herbivore tooth (3·6<C/N)
Herbivore tooth (3·2≤C/N≤3·6)
Herbivore bone + horse tooth (3·6<C/N)
Figure 5. ä13C and ä15N values in Kent’s Cavern mammal tooth and bone collagen.
336 H. Bocherens et al.
& Mariotti, 1992). The ä13C values of enamel carbon-
ate hydroxylapatite of carnivores, herbivores and rock
matrix reveal separated clusters of points for these
three categories of samples (Figure 4). When compared
to the enamel carbonate hydroxylapatite isotopic com-
positions of samples from French caves (Figure 6), the
ä13C values appear very similar for the diVerent trophic
groups.
The diVerence between the ä13C values of enamel
carbonate hydroxylapatite between herbivores and
carnivores has been attributed to the isotopic diVerence
between lipids and carbohydrates of the food (Krueger
& Sullivan, 1984), although more recent investigations
(Ambrose & Norr, 1993; Tieszen & Fagre, 1993) sup-
port an interpretation of homogenization of dietary
carbon through the citric acid cycle. The increased
content of isotopically depleted lipids in the carnivore
diet would tend to shift the carbonate hydroxylapatite
values in the observed direction. It is of interest that
Lee-Thorp et al. (1989) observed no diVerence between
the absolute ä13C values of enamel carbonate hydroxyl-
apatite in modern herbivores and carnivores.
Another way to assess the preservation of enamel
carbonate hydroxylapatite is to compare the diVerence
of the ä13C values in collagen and in carbonate
hydroxylapatite (Ä13C), which has been related to
trophic level and climatic conditions (Figure 7;
Krueger & Sullivan, 1984; Lee-Thorp et al., 1989;
Bocherens & Mariotti, 1992). These Ä13C values were
8·4&1·4‰ (range: 6·8–9·6‰) for modern European
and Alaskan herbivores and 8·6&0·5‰ (range: 7·8–
9·1‰) for Pleistocene French herbivores. In Kent’s
Cavern herbivores for which both collagen and
carbonate hydroxylapatite ä13C values have been
measured present a range of their Ä13C values
from 9·0 to 10·7‰. In Kent’s Cavern carnivores, the
diVerence is 6·1&0·4‰, rather larger than for
modern carnivores (4·3&1·0‰, calculated from
Lee-Thorp et al., 1989). These Ä13C values are thus
slightly higher in Kent’s Cavern herbivores and carni-
vores relative to their modern and Pleistocene trophic
equivalents. This eVect is mostly due to more positive
carbonate hydroxylapatite ä13C values in Kent’s
Cavern specimens, rather than collagen ä13C values.
It remains to be determined what amount of ex-
change has occurred between the carbonate hydroxyl-
apatite and the exogenous carbon with heavier ä13C
values (Table 5). If this exchange occurred, the spac-
ing between the Ä13C values of Kent’s Cavern herbi-
vores and carnivores is nevertheless similar to that
measured in modern mammals from a single ecologi-
cal environment (Bocherens & Mariotti, 1992). By
comparison with ä13C values of modern mammals of
the same geographic and climatic area, the ä13C
values of Kent’s Cavern mammals retain the dis-
crimination between carnivores and herbivores. The
impact of this potential exchange upon the interpret-
ation of omnivory warrants further studies and the
ongoing analysis of radiocarbon dates from paired
collagen/apatite CO2 will help to resolve the issue of
isotopic exchange in the inorganic phase.
The state of preservation of the oxygen isotopic
values is more diYcult to assess. The ä18O values
range from 21·5 to 27·1‰ (Table 1 and Figure 4).
Excursions of several ml are known in marine car-
bonates deposited during the time period from
50,000–30,000 years bp (Bond et al., 1993). The aver-
age values of ä18O are not significantly diVerent
between horses (ä18O=25·7&1·0‰), woolly rhinoc-
eros (ä18O=25·5&1·6‰), deer (ä18O=25·1&2·1‰)
and hyenas (ä18O=25·1&1·2‰). The ä18O values
measured on samples of rock matrix surrounding the
fossils and on some diagenetic calcite formed within a
horse tooth are slightly higher than those measured
on the fossil specimens, with no overlap (27·2–27·9‰;
Table 5 and Figure 4). The ä18O values measured on
the fossil specimens are lower than those measured
on roughly contemporaneous specimens from Caves
located in Southern France, where they range from
26·3 to 31·0‰ (Figure 8; Bocherens, 1992), and may
be due to the latitudinal eVect on ä18O values in rain
water (Yurtsever & Gat, 1981).
Paleobiological implications
The ä13C values of bone collagen are slightly more
negative in Kent’s Cavern herbivores than in equiva-
lent samples from France (around 1‰; Figure 4). The
ä13C values are significantly diVerent between horses
from Kent’s Cavern from Marillac (Student’s t-test,
P<0·01). This diVerence may be due to variations in
the ä13C values of the plants at the base of the food
webs: the consequence of prevailing environmental
conditions (Tiezsen, 1991). Interestingly, van Klinken
et al. (in press) report more negative ä13C values in
–16
12
0
–24 –17
5
9
8
7
6
4
3
2
1
–23 –22 –20 –19
11
10
d
13C (‰)
d
15
N
 (
‰
)
Kent's Cavern
Marillac
Aldene Cave
Mialet Cave
'
Herbivores
(auroch and hare)
Carnivore
Herbivores
(horse)
Herbivores
  H, Horse
  R, Reindeer
  B, Bison
Carnivores
Carnivores
Herbivores
  H, Horse
  Rh, Rhinoceros
  D, Deer
–18
B
R
HRh
H
D
Figure 6. ä13C and ä15N average (&s.d.) values in Kent’s Cavern
mammal bone collagen compared with those of three French
Upper Pleistocene localities (values for French localities are from
Bocherens, 1992, and Fizet, 1992).
Isotopic Signatures From Pleistocene Cave Fauna 337
Holocene herbivore collagen in England than in
France.
The collagen ä13C values from two deer bones
are more negative than the values from the rest of
the herbivores from Kent’s Cavern and from French
upper Pleistocene localities ("23·3 and "23·4‰). If
these values truly reflect a biological signal, they
carry paleobiological meaning, but it is unknown
whether a unique food source or ecological variables
are causal factors. One ecological factor that causes
such negative values in modern herbivore collagen is
the canopy eVect, due to the recycling of respiratory
CO2 and physiological conditions for the plants un-
der canopy (Tieszen, 1991; van der Merwe & Medina,
1991). However, such an explanation for the more
negative bone collagen ä13C values of deer is unlikely
since the paleobotanical data do not support the
occurrence of a close-canopied forest in the surround-
ings of Kent’s Cavern at the time of deposition of the
‘‘cave-earth’’ layer (Campbell, 1977).
–8
Herbivores
Bear
–24 –14–22 –20 –18 –16 –12 –10
d
13C (‰)
Carnivores
Herbivores
Carnivore
Herbivores
Black bear
Cave bears
Herbivores
Carnivores
South Africa
(Lee-Thorp et al., 1989)
Arctic and temperate area
(Bocherens et al., 1994)
Upper Pleistocene from
Aldene and Mialet
(Bocherens et al., 1994)
Kent's
Cavern
'
c a6·8 ± 1·4
4·3 ± 1·0c a
8·4 ± 1·4 ac
4·3
c a
3·9
ac
c 8·6 ± 0·5 a
ac
6·7 ± 0·7
c a10·1 ± 0·6
c a
6·1 ± 0·4
c a
6·1
Figure 7. Average collagen (‘‘c’’) and carbonate hydroxylapatite (‘‘a’’) ä13C values in modern, Kent’s Cavern and Upper Pleistocene French
localities herbivores, carnivores and bears (bars represent isotopic values’ standard deviation).
–8
32
20
–20 –10
30
28
26
24
22
–18 –16 –14
d
13C (‰)
d
18
O
 (
‰
)
Kent's Caven
H, horses
Rh, Rhinoceros
D, Deer
–12
D
RhH
Aldène
Mialet
Hyenas (Kent's Cavern)
Brown bear (Kent's Cavern)
Undet. carnivore (Aldène)
Cave bears (Mialet)
Cave bears (Aldène)
Artiodactyla and perissodactyla
Carnivora
Figure 8. ä13C and ä18O average (&s.d.) values in Kent’s Cavern mammal enamel carbonate hydroxylapatite compared with those of two
French Upper Pleistocene localities (values for French localities are from Bocherens, 1992).
338 H. Bocherens et al.
The ä15N value of the one sample of brown bear
from Kent’s Cavern collagen was among the most
positive of all carnivores and herbivores combined
(Table 2). Based on this isotopic composition which
are very similar to those of hyenas, the brown bear
from Kent’s Cavern had most probably a carnivorous
diet. In contrast, the isotopic measurements of fossil
cave bears indicated a mostly herbivorous diet for
this species (Bocherens et al., 1994). However, more
carnivorous diets are known for some populations of
modern brown bears (Herrero, 1978).
The large number of hyenas analysed in this study
provides, for the first time, a reference set of data for
upper Pleistocene carnivores in Western Europe that
we can compare to the results obtained on cave bears
of the same age. The cave bear belongs to the order
Carnivora but its diet has been shown to be mostly of
plants, according to the low ä15N values of their bone
collagen (Bocherens et al., 1994). However, the diVer-
ence of the ä13C values in collagen and in carbonate
hydroxylapatite of cave bears (6·7&0·7‰) published
by Bocherens et al. (1994) are similar to those of Kent’s
Cavern hyenas (6·1&0·4‰). Nonetheless, these bears
clearly presented more negative ä13C values in both
collagen and carbonate hydroxylapatite when com-
pared to the hyenas. The cave bear Ä13C values appear
also close to those reported for herbivores by Lee-
Thorp (1989), but those herbivores were from a diVer-
ent climatic zone. True carnivores (such as hyenas)
have collagen ä13C values 1–2‰ more positive than
those of their herbivorous prey, and the Ä13C between
collagen and carbonate hydroxylapatite is about 3‰
less than the Ä13C for grazing herbivores. In these two
species of the order Carnivora, the similar Ä13C values,
regardless of trophic position, points out the futility for
attempts to distinguish between plant and animal
sources in omnivores (e.g. other bears) with these Ä13C
values only. The biochemical composition of the
diet and the digestive physiology of the animal are
obviously additional important factors that may
influence Ä13C values.
Conclusions
Chemical discrimination of herbivores and carnivores
is observed in the organic nitrogen and carbon isotopic
values, as well as in the carbonate hydroxylapatite–
collagen diVerences. Between the absolute isotopic
values of paleodietary indicator, there is an obvious
lack of unambiguous omnivore ‘‘space’’, particularly
for the ä13C and the ä15N, but also in the ä13C of
enamel. Combined isotopic paleoindicators such as
spacings in organic/inorganic carbon isotopes as a
function of absolute organic nitrogen or carbon iso-
topic values (e.g. Figure 9) may help to resolve some of
the diYculties in interpreting omnivorous diets. Full
resolution and utilization of these paleodietary indic-
tors will require not only resolution of any diagenetic
overprinting, but also an understanding of the meta-
bolic pathways that contribute to the fractionation of
isotopes from the diet to the animal’s tissue.
References
Ambrose, S. H. (1991). EVect of diet, climate and physiology on
nitrogen isotope abundances in terrestrial foodwebs. Journal of
Archaeological Science 18, 293–317.
Ambrose, S. H. & DeNiro, M. J. (1986). The isotopic ecology of East
African mammals. Oecologia (Berlin) 69, 395–406.
Ambrose, S. H. & Norr, L. (1993). Experimental evidence for the
relationship of the carbon isotope ratios of whole diet and dietary
protein of those of bone collagen and carbonate. In (J. B. Lambert
& G. Grupe, Eds) Prehistoric Human Bone: Archaeology at the
Molecular Level. Berlin: Springer-Verlag, pp. 1–37.
Anderson, E. (1984). Who’s who in the Pleistocene: a mammalian
bestiary. In (P. S. Martin & R. G. Klein, Eds) Quaternary
Extinctions, A Prehistoric Revolution. Tucson, AZ: University of
Arizona Press, pp. 40–89.
Bada, J. L., Peterson, R. O., Schimmelmann, A. & Hedges, R. E. M.
(1990). Moose teeth as monitors of environmental isotopic param-
eters. Oecologia 82, 102–106.
Barnosky, A. D. (1986). ‘‘Big game’’ extinction caused by late
Pleistocene climatic change: irish elk (Megaloceros giganteus) in
Ireland. Quaternary Research 25, 128–135.
Bocherens, H. (1992). Biogéochimie isotopique (13C, 15N, 18O) et
paléontologie des Vertébrés: applications à l’étude des réseaux
trophiques révolus et des paléoenvironnements. Mémoires des
Sciences de la Terre 92–6, Thèse de Doctorat de l’Université Paris
6, 1–317.
Bocherens, H. & Mariotti, A. (1992). Biogéochimie isotopique du
carbone dans les os de mammifères actuels et fossiles de zones
froides et tempérées. Comptes Rendus d l’Académie des Sciences de
Paris 315, 1147–1153.
Bocherens, H., Fizet, M. & Mariotti, A. (1992). Is collagen from
teeth or bones equivalent for isotopic (13C, 15N) diet investi-
gations? Fifth North American Paleontological Convention
(NAPC V), Chicago. Special Publications Series of the Paleonto-
logical Society 6, 30.
Bocherens, H., Fizet, M. & Mariotti, A. (1994). Diet, physiology and
ecology of fossil mammals as inferred from stable carbon and
nitrogen isotope biogeochemistry: implications for Pleistocene
bears. Palaeogeography, Palaeoclimatology, Palaeoecology 107,
213–225.
12
4
d
13C (collagen)
–18–23
10
11
9
6
5
8
–22 –21 –20 –19
7
D
13
C
 (
ap
at
it
e 
– 
co
ll
ag
en
)
Herbivores (Kent's Cavern)
Hyenas (Kent's Cavern)
Cave bears (Bocherens et al., 1994)
Figure 9. Collagen ä13C values versus Ä13C values (diVerence of
ä13C values in collagen and carbonate hydroxylapatite) in Kent’s
Cavern herbivores, hyenas, and French cave bears (values for cave
bears are from Bocherens et al., 1994).
Isotopic Signatures From Pleistocene Cave Fauna 339
Bocherens, H., Fizet, M., Mariotti, A., Lange-Badré, B.,
Vandermeersch, B., Borel, J. P. & Bellon, G. (1991a). Isotopic
Biogeochemistry (13C, 15N) of fossil vertebrate collagen: impli-
cations for the study of fossil food web including Neandertal Man.
Journal of Human Evolution 20, 481–492.
Bocherens, H., Fizet, M., Mariotti, A., Billiou, D., Bellon, G., Borel,
J. P. & Simone, S. (1991b). Biogéochimie isotopique (13C, 15N, 18O)
et paléoécologie des ours Pléistocènes de la grotte d’Aldène. Bul-
letin du Musée d’Anthropologie Préhistorique de Monaco 34, 29–49.
Bond, G., Broecker, W., Johnsen, S., McManus, J., Labeyrie, L.,
Jouzel, J. & Bonanio, G. (1993). Correlations between climate
record from North Atlantic sediments and Greenland ice. Nature
365, 143–147.
Campbell, J. B. (1977). The Upper Palaeolithic of Britain: A Study of
Man and Nature in the Late Ice Age. Oxford: Clarendon Press.
Chaplin, R. E. (1975). The ecology and behaviour of deer in relation
to their impact on the environment of prehistoric Britain. In (J. G.
Evans, S. Limbrey & H. Cleere, Eds) The EVect of Man on the
Landscape: The Highland Zone. Council for British Archaeology,
Research Report No. 11, pp. 40–42.
Dendaletche, C. (1982). Ondes trophiques et stratégies animales
d’utilisation de l’espace. Acta Biologica Montana 1, 14–29.
DeNiro, M. J. (1985). Post-mortem preservation and alteration of in
vivo bone collagen isotope ratios in relation to paleodietary
reconstruction. Nature 317, 806–809.
Estep, M. L. F. & Vigg, S. (1985). Stable carbon and nitrogen
isotope tracers of trophic dynamics in natural populations and
fisheries of the Lahontan Lake system, Nevada. Canadian Journal
of Fisheries and Aquatic Sciences 42, 1712–1719.
Fizet, M., Mariotti, M., Bocherens, H., Lange-Badré, B.,
Vandermeersch, B., Borel, J. P. & Bellon, G. EVect of diet,
physiology and climate on carbon and nitrogen stable isotopes of
collagen in a late Pleistocene anthropic paleoecosystem (France,
Charente, Marillac). Journal of Archaeological Science (in press).
Fogel, M. L., Tuross, N. & Owsley, D. W. (1989). Nitrogen isotope
tracers of human lactation in modern and archeological popu-
lations. Annual Report of the Director of the Geophysical Labora-
tory, Carnegie Institution, Washington, 1988–1989, pp. 111–117.
Garrod, D. A. E. (1926). The Upper Palaeolithic Age in Britain.
Oxford: Clarendon Press.
Guthrie, R. D. (1982). Mammals of the mammoth steppe as paleo-
environmental indicators. In (D. M. Hopkins, J. V. Matthews, Jr.,
C. E. Schweger & S. B. Young, Eds) Paleoecology of Beringia.
London: Academic Press, pp. 307–326.
Heaton, T. H. E., Vogel, J. C., Chevallerie, G. v. l. & Collett, G.
(1986). Climatic influence on the isotopic composition of bone
nitrogen. Nature 322, 822–824.
Herrero, S. (1978). A comparison of some features of the evolution,
ecology and behavior of black and grizzly/brown bears. Carnivore
1, 7–17.
Koch, P. L., Behrensmeyer, A. K., Tuross, N. & Fogel, M. L. (1990).
Isotopic fidelity during bone weathering and burial. Annual Report
of the Director of the Geophysical Laboratory, Carnegie Institution,
Washington, 1989–1990, pp. 105–110.
Koch, P. L., Fisher, D. C. & Dettman, D. (1989). Oxygen isotope
variation in the tusks of extinct proboscidians: a measure of season
of death and seasonality. Geology 17, 515–519.
Krueger, H. W. & Sullivan, C. H. (1984). Models for carbon isotope
fractionation between diet and bone. ACS Symposium Series 258,
205–220.
Lee-Thorp, J. A. (1989). Stable carbon isotopes in deep time. The diets
of fossil fauna and hominids. Ph.D. Thesis, University of Cape
Town.
Lee-Thorp, J. A., Sealy, J. C. & Van der Merwe, N. J. (1989). Stable
carbon isotope ratio diVerences between bone collagen and bone
apatite, and their relationship to diet. Journal of Archaeological
Science 16, 585–599.
Schoeninger, M. J. (1985). Trophic level eVects on 15N/14N and
13C/12C ratios in bone collagen and strontium levels in bone
mineral. Journal of Human Evolution 14, 515–525.
Schoeninger, M. J. & DeNiro, M. J. (1984). Nitrogen and carbon
isotopic composition of bone collagen from marine and terrestrial
animals. Geochimica et Cosmochimica Acta 48, 625–639.
Schwarcz, H. P. (1991). Some theoretical aspects of isotope paleodiet
studies. Journal of Archaeological Science 18, 261–275.
Sealy, J. C., van der Merwe, N. J., Lee-Thorp, J. A. & Lanham, J. L.
(1987). Nitrogen isotopic ecology in southern Africa: implications
for environmental and dietary tracing. Geochimica et Cosmo-
chimica Acta 51, 2707–2717.
SutcliVe, A. J. (1985). On the Track of Ice Age Mammals. London:
British Museum.
Tauber, H. (1986). Analysis of stable isotopes in prehistoric popu-
lations. Mitteilungen der Berliner Gesellschaft für Anthropologie,
Ethnologie und Urgeschichte 7, 31–38.
Tieszen, L. L. (1991). Natural variations in the carbon isotope values
of plants: implications for archaeology, ecology and paleoecology.
Journal of Archaeological Science 18, 227–248.
Tieszen, L. L. & Fagre, T. (1993). EVect of diet quality and
composition on the isotopic composition of respiratory CO2 bone
collagen, bioapatite, and soft tissues. In (J. B. Lambert &
G. Grupe, Eds) Prehistoric Human Bone: Archaeology at the
Molecular Level. Berlin: Springer-Verlag, pp. 121–155.
Tuross, N. & Fogel, M. L. F. (1994). Stable isotope analysis and
subsistence patterns at the Sully Site. In (D. Owsley & R. Jantz,
Eds) Skeletal Biology of the Great Plains. Washington, DC:
Smithsonian Institution, pp. 283–290.
Tuross, N., Fogel, M. L. F. & Hare, P. E. (1988). Variability in the
preservation of the isotopic composition of collagen from fossil
bone. Geochimica et Cosmochimica Acta 52, 929–935.
Tuross, N., Fogel, M. L., Newson, L. & Doran, G. H. (1994).
Subsistence in the Florida archaic: stable isotope and archeo-
botanical evidence from the Windover Site. American Antiquity 59,
288–303.
Tuross, N. & Stathopolos, L. (1993). Ancient proteins in fossil
bones. Methods in Enzymology 224, 121–129.
Van der Merwe, N. J. (1989). Natural variation in 13C concentration
and its eVect on environmental reconstruction using 13C/12C ratios
in animal bones. In (T. D. Price, Ed.) The Chemistry of Prehistoric
Human Bone. Cambridge: Cambridge University Press, pp. 105–
125.
Van der Merwe, N. J. & Medina, E. (1991). The canopy eVect,
carbon isotope ratios and foodwebs in Amazonia. Journal of
Archaeological Science 18, 249–259.
Van Klinken, G. J., van der Plicht, H. & Hedges, R. E. M. Bone
13C/12C ratios reflect (paleo-)climatic variations. Geophysical
Research Letter (in press).
Vereshchagin, N. K. Baryshnikov, G. F. (1982). Paleoecology of the
mammoth fauna in the eurasian arctic. In (D. M. Hopkins, J. V.
Matthews, Jr., C. E. Schweger & S. B. Young, Eds) Paleoecology
of Beringia. London: Academic Press, pp. 267–279.
Vereshchagin, N. K. & Baryshnikov, G. F. (1984). Quaternary
mammalian extinctions in Northern Eurasia. In (P. S. Martin &
R. G. Klein, Eds) Quaternary Extinctions, A Prehistoric
Revolution. Tucson, AZ: University of Arizona Press, pp. 483–516.
Yurtsever, Y. & Gat, J. R. (1981). Atmospheric waters. In (J. R. Gat
& R. Gonfiantini, Eds) Stable Isotopic Hydrology: Deuterium and
Oxygen-18 in the Water Cycle. Vienna: IAEA, Technical reports
series, 210, pp. 103–142.
340 H. Bocherens et al.