See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/222348424

A palaeodietary investigation of carbon (13C/12C) and nitrogen (15N/14N) in

human and faunal bones from the Copper Age cemeteries of Varna I and
Durankulak, Bulgaria

Article in Journal of Archaeological Science · November 2006

DOI: 10.1016/j.jas.2006.02.002

CITATIONS READS

72 193

5 authors, including:

John Chapman Bisserka Gaydarska

Durham University Durham University
20 PUBLICATIONS 166 CITATIONS 89 PUBLICATIONS 1,317 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

The Times of Their Lives View project

Early urbanism in prehistoric Europe?: the case of the Tripillia mega-sites View project

All content following this page was uploaded by Bisserka Gaydarska on 12 November 2019.

The user has requested enhancement of the downloaded file.

 Journal of Archaeological Science 33 (2006) 1493e1504
http://www.elsevier.com/locate/jas

A palaeodietary investigation of carbon (13C/12C) and nitrogen (15N/14N)

in human and faunal bones from the Copper Age cemeteries of
Varna I and Durankulak, Bulgaria
Noah V. Honch a,*, T.F.G. Higham a, J. Chapman b, B. Gaydarska b, R.E.M. Hedges a
a
Research Laboratory for Archaeology and the History of Art, University of Oxford, Oxford OX1 3QY, UK
b
Department of Archaeology, University of Durham, Durham DH1 3LE, UK
Received 10 October 2005; received in revised form 18 January 2006; accepted 1 February 2006

Abstract

Stable isotope analyses have been applied to human and faunal bone collagen from the Varna I and Durankulak cemeteries to explore palae-
odietary adaptations in the Neolithic and Eneolithic (Copper Age). The results suggest both populations primarily utilised terrestrial, C3-based
diets, despite their proximity to the Black Sea. The wider d15N range of the Durankulak humans likely indicates the differential utilisation of
terrestrial meat sources, which is probably related to the degree to which primary and/or secondary ovicaprid products were consumed, partic-
ularly since ovicaprid d15N values differ from other herbivores. The isotopic distribution of Varna I reflects a linear relationship between d15N
and d13C, suggesting that a minority of individuals enriched in both isotopic parameters supplemented their diets with marine resources. These
burials include the well known ‘chieftain’ (burial 43) and show notable material wealth by way of grave goods. At the population level, however,
there is no significant correlation between stable isotope values and material wealth at Varna I, a fact with implications for theories regarding
emergent social/economic hierarchies in Balkan prehistory. Five burials at Durankulak were found to have relatively enriched d13C and d15N
values with respect to the rest of the population. These burials reflect a prominently marine-based or mixed terrestrial C3-based diet that included
C4 inputs, possibly from millet, for which the limitations of stable isotope analysis on bulk collagen are not able to differentiate. AMS dating has
shown that these burials belong to a much later period.
Ó 2006 Elsevier Ltd. All rights reserved.

Keywords: Stable isotope analysis; Palaeodiet; Carbon; Nitrogen; Bulgaria; Neolithic; Eneolithic

1. Introduction East Balkans [53]. To some, Varna has even been the focus
of a discourse related to the birth of European Civilization
The accidental discovery of the Varna I cemetery (hence- (e.g. [32]). Whether or not these connections are tenable,
forth referred to as Varna) over 30 years ago revealed an aston- Varna most certainly represents the earliest known mass con-
ishing assemblage of goldwork, the sheer quantity and centration of gold artefacts in the world [14,26,43,44].
elegance of which rivalled prominent but much later finds in Unlike most Eneolithic (Copper Age) cemeteries in the
Mesopotamia and Egypt [14,43,44]. The volume and diversity East Balkans, Varna is not associated with a settlement. Thir-
of material culture at Varna has widely been interpreted as teen pile dwellings have been located near the Varna Lakes,
proof of social complexity [43], statehood [42], and the exis- which are allegedly coeval with the cemetery [32], but their
tence of widespread inter-regional exchange networks in the relevance to the site is not firmly established. It is therefore
difficult to link the Varna burials to a domestic or agricultural
context that may provide direct archaeological evidence of the
* Corresponding author. Tel.: þ44 1865 285209; fax: þ44 1865 285220. regional diets and subsistence strategies of humans in antiq-
E-mail address: noah.honch@lincoln.ox.ac.uk (N.V. Honch). uity. The situation is compounded by a lack of comprehensive

0305-4403/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jas.2006.02.002
1494 N.V. Honch et al. / Journal of Archaeological Science 33 (2006) 1493e1504

excavation reports, which has posed a formidable barrier to 1.1. The archaeology of Varna
archaeological analysis [12]. Stable isotope analysis is a viable,
direct method of investigating human and faunal diets from The Varna cemetery is located in north east Bulgaria,
preserved biological remains [2,30,41,46]. In light of evidence slightly inland of the Black Sea and north of the Varna Lakes
suggesting the emergence of social differentiation during the (Fig. 1). Archaeological excavations have uncovered approxi-
Eneolithic (see below), we aimed to investigate whether stable mately 280 burials, which are generally classified into three
isotope analysis could shed light on the potential relationship, groups: (i) inhumations (155 burials), (ii) cenotaphs (56
if any, between the diet and status of individuals in the East burials), and (iii) disturbed burials (70 burials) [12,13,32].
Balkans during this period. The number and variety of mortuary goods amongst inhuma-
To the north of Varna is the Eneolithic cemetery of Duran- tions varies greatly. Many inhumations are sparse, accom-
kulak, an impressive site that contains over 1000 inhumations, panied only by pottery fragments, whereas others are
making it one of the largest concentrations of prehistoric incredibly abundant. Burial 43, for instance, contained 990
burials in Europe [51]. The cemetery is associated with gold artefacts that together weighed 1.5 kg. Amongst other
a tell settlement that exists as an island in the western side items, the buried human was associated with numerous gold
of Lake Durankulak, a coastal lagoon with a continuous sedi- necklaces, earrings, appliqués, bracelets, Spondylus bangles,
ment history dating back to the Neolithic [6,7,36]. While the a copper adze with gold rings, and a gold penis sheath. In light
Neolithic settlement on the lake-shore is mostly comprised of its material wealth, burial 43 has been characterised as
of pits and the plentiful deposition of artefacts, the Eneolithic everything from a chieftain to a king [32]. Regardless of this
settlement phases of the tell contain stone-walled architecture individual’s precise status, the material culture is exceptional
and other novel phenomena that are hallmarks of the late even by Varna standards.
Hamangia and Varna cultures and evidence of settlement dif- Cenotaph burials lack human remains, the absence of
ferentiation [5,13,52,54]. The Durankulak tell site is the only which cannot be explained by differential weathering, post-
extensively excavated Eneolithic settlement on the Black Sea depositional decay or physical disruptiondaspects clearly
littoral and provides useful domestic information that may present in disturbed burials [32,44]. Cenotaphs are inferred
aid in the reconstruction of human dietary trends during the to be a deliberate and symbolic act of reverence for individuals
Neolithic and Eneolithic in this region. who died away from the Varna cemetery [13]. In most in-
The aims of this study are twofold: (1) to obtain stable iso- stances, rings, diadems, beads, bracelets, and other decorative
tope data that comparatively assesses the diets of Eneolithic items are placed in anatomical positions on clay masks
humans at Varna and the Neolithic and Eneolithic humans at [43,44]. Overall, this group of burials consistently contains
Durankulak, providing the first evidence of this kind for the the largest, most diverse collections of grave goods.
East Balkans; and (2) to explore the ability of stable isotope The central cluster of Varna burials is referred to as ‘‘the
analysis to relate dietary trends to the inferred social status core’’, which contains a group of symbolic and particularly
of prehistoric humans. wealthy inhumations, including burial 43 and a number of

ROMANIA

Durankulak

Varna
BLACK
SEA

BULGARIA

TURKEY

GREECE

Fig. 1. Location of Varna and Durankulak, Bulgaria.

 N.V. Honch et al. / Journal of Archaeological Science 33 (2006) 1493e1504 1495

cenotaphs [18]. Twelve of the most abundant burials are spectrometer). Conventional replicate measurement errors of
located here, all of which contain bone figurines, Dentalium nylon standards were 0.1& for d13C and 0.3 for d15N.
shells, and exceptionally long flint ‘‘superblades’’ [11]. The
‘‘core’’ lacks female or child burials and may be interpreted 2.1. Bone condition and sample selection
as a strong case of male gender bias [11]. Some burials outside
the ‘‘core’’ do not contain grave goods, although most do, and Many skeletal remains were in poor condition at Varna due
children and females are represented. Despite the lack of to acidic soil conditions. Their friable condition caused prob-
stratigraphic information at Varna, the recent AMS dating of lems in the definition of grave floors at the time of excavation
human and faunal bones suggest it was a single-period burial and frequently rendered age and sex determination problem-
site that was in use for approximately 150 years in the mid- atic. At Durankulak, most excavated burials were intact, yet
fifth millennium BCE [26]. Isotope samples were taken from many bones were fragile due to wet preservation conditions,
a variety of spatial contexts to ensure they were representative causing some to fragment upon removal.
of the population. Most samples from Varna and Durankulak were taken as
whole bone. In the absence of long bones, which were prefer-
entially sampled, identifiable non-long bone elements were
1.2. The archaeology of Durankulak
given priority. Highly fragmentary bones were sampled only
when necessary. In such cases, fragments were chosen on
The Durankulak tell settlement is located on an island in
the expected probability of yielding collagen.
the western side of Lake Durankulak, a large lagoon separated
Palaeodietary reconstructions function on the premise that
from the Black Sea by a 100e200 m strip of sand in northeast-
isotopic signatures of food sources are passed along food
ern Bulgaria (Fig. 1). Stonewalled architecture is used during
chains and register in consumer tissues [49]. It is therefore
the early and late Eneolithicda characteristic phenomenon of
essential for human stable isotope values to be interpreted
the late Hamangia culture and evidence of settlement differen-
with reference to potential food sources (e.g. plants and
tiation [13]. The Neolithic and Eneolithic cemetery is located
animals). The bone collections at the Bulgarian Academy of
on the mainland, approximately 400 m west of the Eneolithic
Sciences, Sofia, and the Regional Museum of History, Varna,
tell, near the Neolithic settlement. AMS radiocarbon dates
were large but heavily biased toward human remains. As
from a variety of Hamangia- and Varna-age burials suggest
such, a supplementary collection of unpublished faunal bones
the Durankulak cemetery was in use for at least 500 years,
from the Durankulak tell settlement was sampled at the
from the early- to mid-fifth millennium BCE [28]. Excavations
Dobritch Historical Museum, Bulgaria. The bones were cata-
have uncovered over 1200 burials, details of which have been
logued, well preserved, and most were taxonomically identifi-
published in full [51]. The Durankulak burials vary in the
able; this collection served as the regional baseline of faunal
abundance and scope of grave goods, including pottery,
isotope values. While the additional samples were representa-
stone-based materials, and metal, though not to the same
tive of the most abundant faunal species in the Durankulak
extent as Varna. Over 800 skeletons have undergone physical
cemetery (Equus asinus hydruntinus, Bos sp., and ovicaprids),
anthropological investigation at the Bulgarian Academy of
they lacked a number of species that were reported to have
Sciences, Sofia [60].
been found at the settlement, including the European hare
(Lepus eurpaeus), the common dolphin (Delphinus delphis),
2. Materials and methods the common porpoise (Phocoena phocoena), wild boar (Sus
scrofa), bear (Ursus arctos), and otter (Lutra lutra), among
Bones were prepared using a modified protocol of Bronk others [50]. A lack of preserved macrofossils precluded the
Ramsey et al. [8]. Briefly, 0.50e0.75 g of bone were shot- analysis of plant remains. Available palynological reports
blasted with aluminium oxide and crushed to a coarse powder from Durankulak lake settlements have provided some
before undergoing demineralisation in 0.5 M HCl for 48 h at evidence of barley (Hordeum sp.), wheat (Triticum sp.), and
<10
C, or until CO2 ceased to evolve. The samples were trace amounts of millet (Panicum miliareum), and lentils
rinsed with deionised MilliQ water, placed in sealed tubes (Lens esculenta) during the late Neolithic and Eneolithic [7].
containing a pH 3 HCl solution and gelatinised at 75
C for
at least 48 h. The supernatant was subsequently filtered with 2.2. Sample integrity
a 5e8 mm EzeeÒ filter (Elkay Laboratory Products) and the
soluble gelatine freeze-dried for 48 h. Insoluble residues We assessed the state of bone preservation using a combina-
were discarded. tion of C:N ratios, collagen yield, and minimum weight yields
Bulk collagen from each sample was weighed in triplicate of carbon and nitrogen. Collagen with a C:N ratio of 2.9e3.4
to between 2.0 and 3.5 mg in tin capsules. Isotopic analyses (akin to modern bone) was believed to yield reliable d13C and
were conducted using an automated carbon and nitrogen d15N values [17,46]. According to Hedges and van Klinken
analyser and a continuous-flow isotope-ratio-monitoring mass [24], bone of ‘‘good preservation’’ contains >20% of the orig-
spectrometer (cf-irm-ms; ANCA Roboprep coupled to a inal collagen (based on a modern value of 200 mg collagen/g
20/20 mass spectrometer or a Carlo Erba carbon and nitrogen of dry bone). Although a high collagen yield cannot ensure
elemental analyser coupled to a Europa Geo 20/20 mass that bone protein is uncontaminated it can improve the
1496 N.V. Honch et al. / Journal of Archaeological Science 33 (2006) 1493e1504

chances of obtaining biogenic signals over bones with little

collagen [55]. Samples were accepted here if they were
Table 2
1% wt% collagen. Minimum weight yields of 0.5 mg for
Stable isotope and grave good data for Varna humans
carbon and 0.2 mg of nitrogen, per combusted sample (2.0e
Sample d13C d15N C:N Measurementsa Wt. % Grave Grave
3.0 mg total weight), were also applied to ensure sample integ-
no. collagen good no. good
rity within the analytical limits of the mass spectrometers. types
Triplicate isotope values were averaged unless otherwise
V8 19.1 10.2 3.2 3 4.7 6 2
stated. Where a triplicate sample measurement did not meet V11 19.3 10.4 3.2 3 9.1 940 7
the minimum analytical thresholds of carbon and nitrogen or V16 19.5 9.2 3.2 3 3.8 1 1
failed due to analytical machinery malfunction, duplicates V17 19.4 9.7 3.2 3 4.9 0 0
were averaged. Single measurements of small yet acceptable V20 19 10.3 3.2 3 2.9 0 0
V25 18.8 10.4 3.2 3 5.1 2 2
collagen were occasionally reported.
V28 19.1 10.9 3.2 3 8 2 2
V30 20 9.1 3.3 2 2.6 7 2
3. Results and discussion V32 19 10.2 3.2 3 7.4 4 3
V34 19.1 10.4 3.2 3 11.5 163 7
V38 19.4 10.2 3.2 3 6.6 0 0
Basic sample descriptions and results of stable isotope anal- V42 18.9 10.1 3.2 3 7.1 1 1
ysis are presented in Tables 1e4. V43 18.5 11 3.2 2 4.8 1013 25
V44 19.1 10.8 3.2 3 5.2 0 0
V45 19.1 10 3.2 3 5.2 6 3
3.1. Animal isotope values
V46 19.4 9.5 3.3 3 3.2 6 5
V47 19.3 9.9 3.2 3 5.9 3 2
The herbivorous mammals (Bos sp., C. elaphus, Equus sp., V50 19.6 10.1 3.2 3 3.2 2 1
ovicaprids) fell within the range of terrestrial C3 consumers V51 18.5 10.9 3.2 3 4.3 16 7
and compared well to Neolithic [31,47], Eneolithic, and later V58 19.1 10.4 3.3 3 4.2 26 2
V67 19.6 10.6 3.2 3 3.6 9 5
Bronze and Iron Age contexts in the Black Sea region [40]
V69 19 10.4 3.2 3 6 2 2
V71 19.7 9.6 3.3 2 2.3 111 9
Table 1 V72 18.8 10.4 3.1 2 6.3 2 2
Faunal stable isotope data from the Durankulak settlement and the Varna V87 19.7 9.1 3.2 1 1.2 3 3
cemetery V94 19.3 10 3.2 2 2.2 1 1
V99 19 9.7 3.2 3 3.4 65 5
Sample Species d13C d15N C:N Measurementsa Wt. % V111 19 11.3 3.2 3 3.9 11 6
collagen V117 19.8 10.2 3.2 3 2.9 9 5
DA25 Badger (Meles meles) 18.3 10.0 3.2 3 10.1 V118 19.1 10 3.3 3 2.1 2 1
D11.2 Bos sp. 20.2 5.9 3.2 3 6.7 V126 19.7 8.7 3.2 3 4.9 5 3
D3u3.1 Bos sp. 19.6 7.2 3.3 3 13.1 V127 19.1 10.2 3.2 3 4.6 2 2
D3u3.2 Bos sp. 19.4 5.9 3.2 3 11.1 V129 19.6 9.7 3.2 3 10.3 1 1
D4.4 Bos sp. 19.8 5.5 3.3 3 13.8 V148 19.5 9.2 3.2 3 4 3 3
D4.6 Bos sp. 19.8 6.3 3.2 3 11.0 V151 19.5 9.9 3.2 3 4.9 13 8
D8a Bos sp. 19.2 5.9 3.2 3 12.2 V158 18.9 9.8 3.3 1 1.8 203 10
DA3 Bos sp. 19.7 5.9 3.2 3 6.7 V160 19 11 3.2 3 3.2 2 1
D4.5 Canid 18.9 8.8 3.2 3 9.1 V171 19.4 10.3 3.3 3 3.5 5 3
DA21 Canid 18.8 10.4 3.2 3 8.2 V174 19.2 9.4 3.2 3 3.2 5 2
DA24 Canid 18.4 8.5 3.2 3 12.5 V179 19.5 9.4 3.3 3 6.7 67 8
D11.1 Deer (Cervus elaphus) 20.1 5.3 3.2 2 6.7 V190 19.3 9.3 3.3 3 3.3 20 6
D3u3.5 Equus sp. 19.5 6.4 3.3 3 7.7 V197 19.3 10 3.2 3 2.6 1 1
D488a Equus asinus hydruntinus 20.6 3.8 3.3 1 2.6 V214 19.8 9 3.4 1 1.3 2 2
DA27 Equus asinus hydruntinus 20.9 3.5 3.2 1 2.6 V215 19.8 10.8 3.2 2 9.5 5 3
DA30 Equus asinus hydruntinus 20.8 3.5 3.2 3 3.4 V225 19.8 9.1 3.2 3 13.3 1 1
DA33 Equus asinus hydruntinus 20.7 3.7 3.2 3 3.0 V234 19.4 9.9 3.2 3 2.6 1 1
DA22 Fox (Vulpes vulpes or 18.5 7.8 3.2 3 17.9 V249 19.4 9.9 3.3 6 6.3 11 4
Vulpes corsac) V251 20 9.2 3.4 2 1.5 5 4
DA23 Fox (Vulpes vulpes or 19.7 8.2 3.2 3 17.0 V252 18.9 10.5 3.2 3 8.5 3 3
Vulpes corsac) V253 19.1 10.4 3.3 3 3.3 10 7
D3u3.3 Sheep (Ovis aries) 19.3 8.1 3.3 3 4.0 V255 18.6 10.4 3.2 2 4.8 5 5
D4.3 Sheep (Ovis aries) 19.2 5.8 3.2 3 11.7 V256 19.5 10.2 3.3 3 5.4 6 3
D971a Ovicaprid 19.4 7.3 3.3 1 3.4 V258 19.2 9.7 3.3 3 4.6 2 2
V33a Ovicaprid 17.7 10.2 3.3 3 4.3 V260 19.7 9.3 3.2 2 1.3 2 2
DA1 Ovicaprid 18.5 10.1 3.2 2 5.8 V265 19.5 10.2 3.2 2 5.8 Unknown Unknown
DA20 Marine turtle 14.1 14.8 3.3 2 4.1 a
Designates the number of isotope measurements that met the criteria of
(unidentified) sample integrity; see section 2.2. Acceptable values are averaged per
a
Designates the number of isotope measurements that met the criteria of individual.
sample integrity; see section 2.2. Acceptable values are averaged per
individual.
 N.V. Honch et al. / Journal of Archaeological Science 33 (2006) 1493e1504 1497

Table 3 data, causative explanations of E. asinus hydruntinus extinc-

Stable isotope data for Durankulak outliers tion are premature; further archaeological and palaeoecologi-
Sample d13C d15N C:N Measurementsa Wt. % Sex cal data are required to address this issue.
no. collagen Most ovicaprids (3e4 individuals) have enriched d15N
D4.2 16.1 11.1 3.2 2 8.8 F values for herbivores, falling within range of dog and fox.
D27 15.5 12.2 3.2 2 5.8 F The observed pattern was not restricted to North East Bulgaria
D32 15.7 10.1 3.2 3 11.2 M
D101 16.1 9.6 3.2 3 17.2 F
but was attested at regional, inland sites in Western Siberia,
D229 16.7 12.3 3.2 3 7.5 Child Neolithic Çatalhöyük in South-Central Turkey, and the north-
a
Designates the number of isotope measurements that met the criteria of
ern Black Sea coast in later periods (Table 5). However, the Du-
sample integrity; See section 2.2. Acceptable values are averaged per rankulak ovicaprid d15N values are depleted by 2& compared
individual. to those at contemporaneous inland, Ukranian sites. Differing
climatic and agronomic conditions may account for the ele-
vated d15N values at Durankulak. Collagen d15N values are
(Fig. 2, Table 5). The mean d15N values of herbivorous fauna, in some cases known to increase with decreasing rainfall, and
excluding ovicaprids, are depleted (3.6e6.4&) in comparison to become depleted with increased temperature/aridity [3,21]
with the omnivores/carnivores (8e10&), reflecting a trophic and resulting water stress [19]. In the broader context, however,
level effect of approximately 3&, a well-attested phenomenon there is little evidence for high temperatures and/or low rainfall
in terrestrial ecosystems (e.g. [37,48]. in the region, especially since other fauna do not appear to be
The d15N value of the Bronze Age turtle is substantially en- affected in this way. The stable isotope values of the remaining
riched over all other samples, attesting to the numerous trophic fauna are within ranges typically observed in temperate, well-
levels in marine environments [44]. The enriched d13C value watered environments. In addition, this would have had to have
(14.1&) is likely the result of dissolved carbonate and bicar- been a long-term process in order to isotopically register in
bonate carbon sources for photosynthesis in marine foodwebs bone collagen, a tissue with a slow turn-over rate [2,34]. It
[33] in the Black Sea. This value is comparable to humans has recently been noted from isotopic studies of hair keratin
(w12 to 13&) and animals (dog: w12 to 14&) who in sheep from Turkey that grazing in contained areas is corre-
consume substantial quantities of marine resources [10,15]. lated with enriched d15N values relative to open, rural contexts,
The d13C (20.7 to 20.1&) and d15N values (3.5e3.8&) which might be due to natural manuring and the increased ni-
for E. asinus hydruntinus are unusually depleted in compari- trification of soils from urine and faeces [23]. It is possible that
son to the remaining terrestrial herbivores. This wild species, the elevated d15N values of the Durankulak ovicaprid collagen
better known as the ‘European wild ass’, inhabited South may similarly be explained as the result of controlled grazing in
Asia and southern regions of Western Europe until it became areas adjacent to or on the tell settlement. Currently, the few re-
extinct in the early Holocene [9,39,50,58]. E. asinus hydrunti- ported values here represent the sum total of d15N values for the
nus persisted for longer periods in the Eastern Balkans and the Black Sea region and the elucidation of ovicaprid d13C and
remains from Durankulak appear to be the latest yet discov- d15N variability requires the analysis of more samples from
ered. E. asinus hydruntinus is found in domestic and burial a variety of environments.
contexts, suggesting it was hunted for food as well as ritual.
Its extinction was probably caused by a combination of eco- 3.2. Human isotope values from Varna and Durankulak
logical and human-induced pressures. Mild climatic condi-
tions and increased forest cover likely reduced the steppe The Varna d13C values range from 20.0 to 18.5&, with
environment in which E. asinus hydruntinus likely thrived a population average of 19.3  0.3& (n ¼ 55), and d15N
[9,20], while intensified hunting during the Neolithic and values ranging from 8.7 to 11.3& with an average of
Eneolithic added further selective pressures [50]. 10.0  0.6&. The vast majority of human d15N are enriched
E. asinus hydruntinus d15N values are approximately one by at least 3& over non-ovicaprid terrestrial herbivores
trophic level below horse, deer, and cattle. In general, horse (w5e6&), exhibiting a trophic level effect and demonstrating
d15N values are comparatively depleted, especially relative the importance of terrestrial meat sources in the diet. The Du-
to domesticated ruminants, and this is ascribed to their diges- rankulak humans show a wider spread in d15N, ranging from
tive physiology. However, here the d15N depletion for E. asi- 7.6 to 11.5&, with a population average of 9.3  0.8&
nus hydruntinus appear especially marked. Considering the (n ¼ 78), while the d13C values closely resemble Varna and
depleted d13C values, E. asinus hydruntinus probably main- range from 19.8 to 18.6&, averaging 19.1  0.3&. Al-
tained a stable, low protein diet of C3 grasses. The distinctive though the d13C values of both populations are slightly en-
isotope values may be because E. asinus hydruntinus exploited riched over humans from inland systems in continental
environmental niches or food resources that other herbivores Europe, which are usually between 20 and 21& (see
did not, by choice or because of hunting pressures. Whatever [56]), the difference is primarily attributable to climatic varia-
the case, it did not experience long-term dietary stress (see tions. Comparative database studies of wood, charcoal, and
[27]) or occupy arid to dry fringe-environments (see [3,19, bone samples from European archaeological sites have shown
21]), which might be expected if heavily hunted, for we would that d13C values tend to become enriched in a north to south
expect to observe highly enriched d15N values. With limited direction, following the climatic/temperature cline that
1498 N.V. Honch et al. / Journal of Archaeological Science 33 (2006) 1493e1504

Table 4 Table 4 (continued )

Stable isotope data for Durankulak humans
Sample d13C (&) d15N (&) C:N Measurementsa Wt. % Sex
Sample d13C (&) d15N (&) C:N Measurementsa Wt. % Sex no. collagen
no. collagen
D811 18.9 8.4 3.2 3 2.4 M
D20 19.1 9.3 3.2 2 1.4 M D812 19.1 8.9 3.2 2 15.7 M
D34 19.1 9.3 3.2 3 3.1 M D813 19.0 7.8 3.2 3 5.3 M
D86 19.4 7.9 3.2 3 2.9 M D815 18.7 9.0 3.2 3 6.8 M
D128 18.8 9.6 3.1 3 3.9 ? D817 19.2 8.2 3.2 4 3.5 M
D193 18.6 10.5 3.2 2 9.3 M D825 18.9 9.7 3.2 3 7.3 M
D450 19.8 9.7 3.2 3 1.6 F D826 19.1 9.5 3.2 3 1.5 ?
D460 19.6 10.0 3.2 3 2.0 F D864 19.2 9.4 3.3 3 1.8 F
D462 18.9 9.3 3.2 3 7.7 F D865 19.0 8.7 3.3 3 4.4 Infant
D465 19.6 9.5 3.2 3 2.4 F D869 19.5 9.2 3.2 1 3.8 M
D483 18.9 9.0 3.2 2 7.6 M D897 19.3 7.9 3.2 3 3.7 F
D484 19.2 8.6 3.2 3 2.6 M D898 19.1 11.2 3.2 2 6.2 F
D485 19.1 8.9 3.2 2 5.8 M D939 19.0 10.1 3.2 3 3.0 M
D488 18.9 10.1 3.2 1 4.0 M D994 19.7 9.6 3.2 1 1.9 M
D491 19.5 8.2 3.2 2 3.9 M D1026 19.0 9.3 3.2 3 5.7 ?
D512 19.6 9.9 3.2 3 2.6 M a
Designates the number of isotope measurements that met the criteria of
D527 19.3 9.8 3.2 3 6.2 F
sample integrity; See section 2.2. Acceptable values are averaged per
D532 19.0 11.5 3.2 3 4.5 F
individual.
D536 19.1 10.1 3.2 1 4.2 M
D553 19.4 10.4 3.2 3 6.6 F
D563 19.2 8.4 3.2 3 2.5 F increases toward the Mediterranean and, apparently, the prox-
D569 18.9 8.0 3.2 3 3.7 F
D588 19.2 10.7 3.2 3 5.5 F
imity to large water bodies like the Black Sea [56,57]. On the
D595 18.6 11.1 3.2 2 3.6 M whole, it is clear that the Varna and Durankulak humans
D596 19.3 10.0 3.2 3 9.2 F primarily utilised terrestrial, C3-based diets with varying
D602 18.9 10.4 3.2 2 5.3 M proportions of meat, particularly in the case of Durankulak.
D606 19.0 10.6 3.2 2 4.2 M It is interesting, however, to note differences in the isotopic
D607 19.3 9.0 3.2 3 8.5 M
D609 19.7 9.7 3.3 3 7.8 M
distributions of Varna and Durankulak (Fig. 3). Whilst the
D614 18.9 9.2 3.2 3 4.4 M d15N vs d13C distribution of Durankulak is roughly circular,
D619 19.2 8.2 3.2 3 4.5 M with no systematic correlation between the observed variables
D621 19.6 10.0 3.2 2 4.2 M (R ¼ 0.012, p ¼ 0.92), the isotope distribution of Varna
D622 19.4 9.8 3.2 2 6.0 M reflects a linear relationship (R ¼ 0.61, p < 0.001) with
D629 19.6 7.6 3.3 4 4.1 M
D630 19.0 8.8 3.2 2 3.3 M
a d15N:d13C gradient of approximately 1:1. Although the
D631 19.5 9.0 3.4 3 3.0 M d13C values are characteristic of terrestrial, C3-based diets,
D635 18.9 9.9 3.2 3 5.0 M the observed d13C vs d15N relationship of Varna is similar to
D636 18.8 9.3 3.2 2 4.5 M that observed in populations where marine protein made sig-
D643 19.3 10.5 3.2 2 3.5 M nificant contributions to the diet [48,59]. Linear trends with
D645 19.6 9.4 3.3 3 3.1 F
D651 18.9 10.0 3.2 3 5.3 M
d15N:d13C gradients of w2:1 have been observed between Eu-
D678 18.8 8.8 3.2 2 8.2 M ropean Neolithic populations, which primarily consumed
D693 18.7 8.3 3.2 1 3.6 M terrestrial food resources, and earlier Mesolithic populations,
D697 18.8 8.6 3.2 2 4.5 F which mostly utilised marine resources (e.g. [35,45]), which
D706 19.3 9.0 3.2 2 6.7 M tends to support basic assumptions about the preferential rout-
D713 18.9 8.2 3.2 3 2.6 ?
D714 19.1 8.4 3.2 2 3.1 F
ing of dietary protein to collagen (as opposed to de novo syn-
D723 19.0 9.0 3.2 3 6.1 M thesis) when the intake of dietary protein is high [41]. It is well
D730 19.2 8.2 3.2 1 10.2 F known that marine food sources are relatively protein-rich and
D731 19.3 11.0 3.2 3 4.0 F have elevated d13C and d15N values with respect to terrestrial
D734 18.6 9.0 3.2 3 7.0 F food [41]. Controlled feeding studies on rats [4] and pigs [30]
D741 19.2 9.2 3.2 1 2.0 F
D747 19.4 10.0 3.2 1 4.6 F
have shown that collagen d13C and d15N values tend to reflect
D751 18.8 9.7 3.2 3 3.5 M those of dietary protein in high-protein diets. Thus, where ma-
D758 19.0 8.8 3.2 3 4.7 M rine diets are utilised in any quantity, it would be expected that
D761 19.1 8.8 3.2 2 3.1 F dietary protein be directly routed to collagen and, depending
D768 18.8 9.0 3.2 1 5.7 M on the degree to which marine foods were consumed, collagen
D770 19.2 8.5 3.2 3 3.1 M
D772 18.9 8.8 3.2 1 8.2 M
should show elevated d13C and d15N with respect to terrestrial,
D773 18.8 8.9 3.2 3 2.6 F C3-based diets. In contrast, although we would expect en-
D777 19.2 8.6 3.3 3 2.7 M riched d15N values for high protein terrestrial diets (with no
D789 19.0 9.0 3.2 3 6.9 M marine protein), we would not anticipate a significant increase
D792 19.0 9.0 3.2 3 3.2 M in d13C compared to terrestrial diets of similar composition but
D794 19.0 9.8 3.2 2 3.7 M
of lower protein intake [22].
 N.V. Honch et al. / Journal of Archaeological Science 33 (2006) 1493e1504 1499

16.0
Varna Humans n=55
14.0 Durankulak Humans n=78
Durankulak Outliers n=5
12.0
fox n=2
10.0 dog n=3

δ15NAIR
badger n=1
8.0
sheep/goat n=5
6.0 cattle n=7
horse n=1
4.0
E. asinus hydruntinus n=4
2.0 deer n=1
BA turtle n=1
0.0
-22.0 -20.0 -18.0 -16.0 -14.0
δ13CVPDB

Fig. 2. Human and animal bone collagen isotope data from Varna and Durankulak, north east Bulgaria. All means expressed as 1s.

The isotope values of humans at Varna reflect a continuum (representing approximately nine individuals or 17% of the
of dietary protein sources ranging from primarily terrestrial C3 sample population; Fig. 4). The latter includes individuals
protein to those that included a combination of mostly terres- with d15N values close to or greater than 11&, especially
trial C3 protein and a detectable component of marine those that have d13C values beyond 1s of the population
resources. The former is exemplified in individuals that are mean (burials 43 and 51). On the whole, however, the con-
depleted in both d15N and d13C beyond the 1s range sumption of marine protein was apparently minor and

Table 5
Comparative stable isotope data from regional and/or contemporary fauna

a
From the sites of Bugor and Bil’shivtsi, c. 500 km north west of the Black Sea.
b
Between the Don and Volga rivers.
Shaded rows denote data collected from this study.
1500 N.V. Honch et al. / Journal of Archaeological Science 33 (2006) 1493e1504

13.0
Varna
y = 1.02x + 29.7
12.0 2 Varna humans
R = 0.37
n=55
11.0
Durankulak
humans n=78

δ15NAIR
10.0
Linear (Durankulak
9.0 humans n=78)
Linear (Varna
8.0 humans n=55)
Durankulak
7.0 y = 0.03x + 9.94
R2 = 1.0x10-4
6.0
-20.5 -20.0 -19.5 -19.0 -18.5 -18.0
δ13CVPDB

Fig. 3. Human stable isotope data for Varna I and Durankulak. Note the relative distribution of each population.

supplemental to the diet. If individuals were consuming signif- goat in addition to terrestrial herbivores and C3 plants. The
icant amounts of marine protein we would expect radiocarbon d13C values of these individuals are close to the population
offsets between co-interred human and terrestrial animals. Re- mean and are consistent with terrestrial C3 protein sources,
cent research suggests there is a 14C reservoir effect in the suggesting their d15N values do not reflect marine protein con-
Black Sea of about 415 years [1]. The AMS dating of select sumption. It is important to note, however, that a small number
(n ¼ 3) humaneterrestrial animal pairs at Varna suggest there of Durankulak humans may have utilised marine resources,
is a low probability of offset from true age due to reservoir ef- such as burials D193 and D595. The isotopic signatures of
fects [26], suggesting little dietary protein was derived from these individuals overlap with the outlying burials at Varna
marine resources, despite the close location of Varna to the (V43 and V51) and exhibit enriched d13C and d15N values.
Black Sea.
There is no significant linear relationship between d15N and 3.3. Isotopes and social stratification at Varna
d C values at Durankulak. The observed spread of d15N and
13

d13C values is likely due to the differential intake of terrestrial, A central objective of this study was to determine whether
C3-based protein. About 23% of the Durankulak humans (18 in- a relationship existed between the diet of specific individuals
dividuals) have lower d15N values than the most depleted individ- at Varna, as inferred from bulk collagen isotope values, and
ual at Varna and are enriched by less than 3& over Bos sp., the their associated material culture, as expressed by the number
most appropriate isotopic benchmark for terrestrial herbivores. and diversity of grave goods, which may be a reflection of
Although terrestrial meat made important contributions to status in prehistory [43,44]. At the population level, the corre-
the diet of these individuals, it did not comprise as much of lation between d13C values and the number or types of grave
the overall diet as it did with individuals that exhibited rela- goods per burial was not statistically significant (Table 6).
tively enriched d15N values, such as burials D532 (11.5&), The relationship between d15N and grave good variables was
D643 (10.5&), and D731 (11.0&). The d15N of the latter also not statistically significant.
are upwards of 2e3& more enriched than ovicaprids, suggest- In specific cases, however, there are correlations between
ing their diets were likely supplemented with terrestrial meat stable isotopes and material culture. The obvious examples
and/or secondary products (e.g. milk, cheese) from sheep/ are burial 43, which contains the largest (1013 artefacts) and

12.5

12.0 Varna Humans n=55

11.5 Burial 28
Burial 43
11.0
Burial 44
δ15NAIR

10.5 Burial 51
Burial 67
10.0
Burial 111
9.5 Burial 160
Burial 215
9.0
Burial 255
8.5 Total Distribution
8.0
-20.5 -20.0 -19.5 -19.0 -18.5 -18.0
δ13CVPDB

Fig. 4. Stable isotope distribution for Varna humans. Select burials are highlighted.
 N.V. Honch et al. / Journal of Archaeological Science 33 (2006) 1493e1504 1501

Table 6 clustering of graves and the differential allocation of grave

Linear Regression Results of Stable Isotope Data vs Grave Good Variables goods, we attempted to determine whether stable isotope
Variables R p values were related to the spatial patterning of burials. The dis-
13
d C vs Grave Good Number 0.241 0.079 tribution of stable isotopes showed significant overlaps
d13C vs Grave Good Types 0.263 0.055 between burial clusters, suggesting the overarching diet of
d15N vs Grave Good Number 0.218 0.113 each group was quite similar (Fig. 5). It is interesting to
d15N vs Grave Good Types 0.196 0.155
note, however, that the population mean of the ‘‘core’’ (area
E) is, in a similar way burials 43 and 51 are to the rest of
most diverse (25 types) collection of artefacts at Varna, and the population, enriched in both d13C and d15N relative to
burial 51, which reflects moderate material wealth (16 arte- the other sample groups. The ‘‘core’’ of the Varna cemetery
facts of seven different types). Burials 43 and 51 are the contains most of the materially wealthy inhumations (but
only to show substantial enrichments in both d13C (>2s above also a minority of materially poor ones), including the vast
the mean) and d15N (Fig. 4). In these cases, the relationship majority of cenotaph burials. Again, a subtle relationship
between the isotope signatures and the material wealth of may be reflected between the diet and inferred status of this
the individuals could be a reflection of their status in antiquity. group in the past.
In the context of the previous discussion, the humans in burials It is clear that no direct and unambiguous connections can be
43 and 51 likely consumed more marine protein than the rest drawn between diet, as inferred from stable isotope data, and sta-
of the population, suggesting long-term dietary differences do tus, as inferred from the material wealth of burials at Varna and
correlate with the material wealth and inferred status in special Durankulak. This may be due not only to the limited resolution
cases. These exceptions, however, do not contravene the find- of stable isotope data but also to the theoretical complexities of
ing that most materially rich burials (burials 11, 34, 58, 71, statusdbe it social, political, religious, or economicdand its
99, 151, 158, 179, 190, 249, 253) have d13C and d15N values assignation to prehistoric humans on the basis of material re-
that are much closer to the population mean (see Table 2). It mains. Although individuals must have maintained a sufficient
is also clear that the isotopes of the unique Varna burials over- local connection to have warranted their burial at Varna or Du-
lap with individuals at Durankulk (Fig. 3). Interestingly, these rankulak, it is possible that some of the populations were origi-
Durankulak burials lack material wealth; D193 contained the nally from distant regions or were local in origin but travelled to
fragments of a ceramic vessel, as did D595 in addition to other regions throughout their lives (e.g. individuals of promi-
E. asinus hydruntinus teeth in the grave fill [51]. These incon- nent economic/political standing or long distance specialists
sistencies are difficult to interpret despite the relative lack of sensu Helms [25]). Some of the observed isotope variation at
disparity in material wealth between most Durankulak burials. Varna and Durankulak may thus be due to the consumption of
It may be that the unique Durankulak individuals consumed the food from a variety of geographic regions or the exposure to cli-
same diet as the Varna outliers or simply consumed more terres- matic or environmental influences that were different from the
trial meat than most at Durankulak and the bulk isotope data local area. This might especially be expected at Varna, which
cannot differentiate between the two. In either case, the Duran- is not affiliated with a settlement. However, it is the Durankulak
kulak data cannot be explained on the basis of inferred status. burialsdwhich are associated with a settlementdthat show
greater isotopic variability. The Durankulak cemetery is much
3.4. Stable isotopes and the spatial clustering of burials larger and was in use for much longer than Varna; it is possible
at Varna that the observed isotopic variability is the result of dietary
changes that are not currently apparent due to the limitation of
Recent chronometric evidence suggests the Varna cemetery the few radiocarbon dates (19) to separate the isotope data
was a single-phase site [26,28]. In light of the spatial into smaller temporal units.

11.5 Area E n=12

Area EC n=1
11.0
Area N n=6
10.5 Area NE n=13
δ15NAIR

10.0 Area W n=4

Area WC n=18
9.5
Burial 43 (E)
9.0 Burial 255 (NE)

8.5 Burial 51 (WC)

Burial 28 (WC)
8.0
Burial 44 (E)
-20.5 -20.0 -19.5 -19.0 -18.5 -18.0
Burial 160 (N)
δ13C VPDB

Fig. 5. Varna human isotope data with respect to burial distribution. E, east; NE, north east; WC, west central; N, north.
1502 N.V. Honch et al. / Journal of Archaeological Science 33 (2006) 1493e1504

3.5. Isotopic outliers at Durankulak 3.6. Sex-based considerations

Five individuals at Durankulak (D4.2, D27, D32, D101, There was close similarity between the stable isotope values
D229) showed unusually enriched d13C values (w4&) com- of the male and female individuals at Durankulak (female:
pared to the rest of the population (Fig. 2). AMS results indi- d13Cmean ¼ 19.3  0.3&, d15Nmean ¼ 9.5  1.0&, n ¼ 25;
cate that three of the outliers (D4.2, D27, D32) date to a much male: d13Cmean ¼ 19.1  0.3&, d15Nmean ¼ 9.2  0.8&,
later period (w1000 14C BP; see [28,29]) than the rest of the n ¼ 48), suggesting there were no obvious, sex-based differ-
cemetery. The dated outliers represent the only group of late ences detectable in the overarching diet of these groups. The
burials at the site and are likely contemporaneous with the un- sexes of Varna humans have not yet been assigned.
dated samples (D101, D229) in light of their isotopic similar-
ity. The systematic d13C enrichment of these individuals may 3.7. Lack of diachronic changes in diet at Durankulak
be attributed to one factor or a combination of factors. The
collective d13C and d15N data (d13Cmean ¼ 16.0  0.4&, Todorova [51] has assigned a relative sequence to the Duran-
d15Nmean ¼ 10.9  1.1&, n ¼ 5) suggest marine resources kulak burials on the basis of their material culture. We evaluated
made significant contributions to the diet, more so than in our stable isotope results with reference to each group. The iso-
the Neolithic and Eneolithic periods. tope values were indistinguishable from each other (Table 7),
Millet consumption, however, cannot be ruled out. Millet suggesting little to no dietary change between the Neolithic
is, in essence, the only widespread C4 plant in Europe that and Eneolithic periods. It should be noted, however, that AMS
has contributed to the human diet, which could lead to en- radiocarbon dating has uncovered several problems with this ty-
riched d13C bone collagen values. According to archaeological pological framework, making firm conclusions premature [28].
and palaeodietary studies, millet was an important component
of the human diet by the Late Bronze or Iron Age [38], and 4. Conclusions
palynological evidence suggests small amounts of millet
may have been present at Durnankulak as early as the Neo- Stable isotope data from Varna and Durankulak suggest hu-
lithic [7]. Since the d15N signatures of the medieval Duranku- mans at both sites maintained terrestrial, C3-based diets where
lak outliers (d15Nrange: 9.6e12.3&) are only marginally protein was mostly derived from terrestrial sources, with a pre-
enriched over the Eneolithic population (d15Naverage: 9.3&), dominance of animal products. The correlated distribution of
they could have maintained a mixed terrestrial/marine diet isotopes at Varna, however, is comparable to populations
where a portion of the terrestrial diet included millet. At the where marine resources made measurable contributions to
level of bulk collagen, however, the influence of C4 plants the diet. The isotopic data are consistent with a minority of
and marine resources cannot be resolved. Compound specific the population utilising some marine resources. Burials 43
stable carbon isotope analysis of single amino acids (e.g. and 51 possessed substantial material wealth compared with
[16]) may be able to resolve this problem. all other burials as well as exhibited enriched d15N and d13C
On the basis of (1) relatively young AMS radiocarbon values. The isotopic data from Durankulak suggest the popu-
dates and (2) strikingly different stable isotope data to the lation utilised an almost exclusive terrestrial, C3-based diet
rest of the population, the burials in question appear to be re- with differential inputs of terrestrial meat sources. Individuals
cent. It is clear, however, that some contain Hamangia-age with relatively enriched d15N values probably consumed pri-
material culture [51]. There are two possible scenarios to ex- mary or secondary ovicaprid products in addition to other ter-
plain this dilemma. First, in a small number of medieval restrial resources whilst those with lower d15N mostly
cases, the mourners placed a few examples of much earlier consumed non-ovicaprid protein sources.
material culture, which was abundant on or near the surface Despite correlations between stable isotope data and mate-
of the medieval settlement, into the graves of their own rial culture in select cases (e.g. burials 43 and 51 vs other Varna
newly dead, perhaps to memorialise the associations of the burials; the ‘‘core’’ vs other areas at Varna), there is no convinc-
deceased with the long history of their village. The second ing evidence for an overarching trend involving diet and status
possibility is that the Copper Age sherds were accidentally at the population level at Varna. Although this does not refute
incorporated into the medieval grave through the digging of the possible wealth and social stature of certain individuals in
the grave-pit. the Chalcolithic, it implies that long-term dietary trends and
differential access to unique or isotopically distinct food re-
sources were not strongly connected with such positions. In
Table 7 any case, it is important to note that some of the observed iso-
Stable isotope data by temporal grouping of human burials, Durankulak topic variation at Varna may be due to environmental influ-
Temporal Group n d13C (&) d15N (&) ences associated with residential differences in the buried
Hamangia I-II 25 19.1  0.2 9.4  0.8 population. Without a firm connection to a settlement, the bur-
Hamangia II-III 1 18.6 10.5 ied population may have included individuals with non-local
Hamangia III 37 19.1  0.3 8.9  0.7 diets, particularly since collagen turns over slowly.
Hamangia IV 11 19.4  0.3 10.0  0.7 All five isotopic outliers at Durankulak are intrusive to the
Varna Age 3 19.2  0.1 9.9  0.8
cemetery and had diets that were considerably different from
 N.V. Honch et al. / Journal of Archaeological Science 33 (2006) 1493e1504 1503

the rest of the population. The enriched d13C values indicate [11] J. Chapman, Colourful prehistories: the problem with the Berlin and Kay
the consumption of marine resources, but it is possible that Colour Paradigm, in: A. Jones, G. MacGregor (Eds.), Colouring the Past,
Berg, Oxford, 2002, pp. 45e72.
millet was also an important part of their diet. Stable isotope [12] J. Chapman, Fragmentation in Archaeology: People, Places and Broken
evidence suggests there were little to no diachronic or sex- Objects in the Prehistory of South Eastern Europe, Routledge, London,
based differences in diet at Durankulak. 2000.
[13] J. Chapman, The creation of social arenas in the Neolithic and Copper
Age of S.E. Europe: the case of Varna, in: P. Garwood, D. Jennings,
Acknowledgements R. Skeats, J. Toms (Eds.), Sacred and Profane: Proceedings of a Confer-
ence on Archaeology, Ritual, and Religion, Oxford University Press,
Oxford, 1991, pp. 152e171.
The authors would like to thank H. Todorova, V. Slavchev, [14] J. Chapman, Social inequality on Bulgarian tells and the Varna problem,
B. Dimitrova, and Y. Yordanov for their collaboration on this in: R. Samson (Ed.), The Social Archaeology of Houses, Edinburgh
project. The research underpinning this article formed part University Press, Edinburgh, 1990, pp. 49e92.
of Noah Honch’s MSc at the University of Oxford. John Chap- [15] B.S. Chisholm, D.E. Nelson, H.P. Schwarcz, Stable-carbon isotope ratios
as a measure of marine versus terrestrial protein in ancient diets, Science
man and Bisserka Gaydarska acknowledge the financial
216 (1982) 1131e1132.
support of the British Academy for travel funding; John Chap- [16] L.T. Corr, J.C. Sealy, M.C. Horton, R.P. Evershed, A novel marine die-
man thanks the Bulgarian Academy of Science for travel and tary indicator utilising compound-specific bone collagen amino acid
subsistence within Bulgaria. Noah Honch acknowledges d13C values of ancient humans, Journal of Archaeological Science 32
Lincoln College, Oxford, for his travel and subsistence costs. (2005) 321e330.
[17] M.J. DeNiro, Postmortem preservation and alteration of in vivo bone
We are grateful to Todor Dimov for providing access to the
collagen isotope ratios in relation to palaeodietary reconstruction, Nature
animal bones at the Historical Museum of Dobritch, and to 317 (1985) 806e809.
Sonia Dimitrova for her helpful assistance. We would also [18] A. Fol, J. Lichardus, Macht, Herrschaft und Gold, Moderne Galerie des
like to thank Nikolai Spassov for access to the Durankulak Saarland-Museums, Saarbrücken (1988).
faunal collection at the National Museum of Natural History [19] D.R. Gröcke, H. Bocherens, A. Mariotti, Annual rainfall and nitrogen-
isotope correlation in macropod collagen: application as a palaeopreci-
and to Kalin Dimitrov for drilling the animal bones.
pitation indicator, Earth and Planetary Science Letters 153 (1997)
279e285.
[20] V. Gromova, Istoriya loshadei (Gen. Equus) v starom svete II, Evolutsiya
References I klasifikatsiya roda, Moskva, Leningrad, 1949.
[21] T.H.E. Heaton, J.C. Vogel, G. von al Chevallerie, G. Collett, Climatic in-
[1] A.E. Aksu, R.N. Hiscott, D. Yasxar, F.I. Isxler, S. Marsh, Seismic stratigra- fluence on the isotopic composition of bone collagen, Nature 322 (1986)
phy of Late Quaternary deposits from the southwestern Black Sea shelf: 822e823.
evidence for non-catastrophic variations in sea-level during the last [22] R.E.M. Hedges, Isotopes and red herrings: comments on Milner et al. and
w10000 yr, Marine Geology 190 (2002) 61e94. Lidén et al, Antiquity 78 (2004) 34e37.
[2] S.H. Ambrose, Isotopic analysis of palaeodiets: methodological and [23] R.E.M. Hedges, J.M.A. Thompson, B.O. Hull, Stable isotope variation in
interpretive considerations, in: M.K. Sandford (Ed.), Investigations of wool as a means to establish Turkish carpet provenance, Rapid Commu-
Ancient Human Tissue, Gordon and Breach Science Publishers, nications in Mass Spectrometry 19 (2005) 3187e3191.
Langhorne, Pennsylvania, 1993, pp. 59e130. [24] R.E.M. Hedges, G.J. van Klinken, A review of current approaches to the
[3] S.H. Ambrose, Effects of diet, climate and physiology on nitrogen pre-treatment of bone for radiocarbon dating, Radiocarbon 34 (1992)
isotope abundances in terrestrial foodwebs, Journal of Archaeological 279e291.
Science 18 (1991) 293e317. [25] M.W. Helms, Crafts and the Kingly Ideal: Art, Trade, and Power, Univer-
[4] S.H. Ambrose, L. Norr, Experimental evidence for the relationship of the sity of Texas Press, Austin, 1993.
carbon isotope ratios of whole diet and dietary protein to those of bone [26] T. Higham, J. Chapman, V. Slavchev, B. Gaydarska, N.V. Honch, C.
collagen and carbonate, in: J.B. Lambert, G. Grupe (Eds.), Prehistoric Bronk Ramsey, Y. Yordanov, B. Dimitrova, New AMS radiocarbon dates
Human Bone: Archaeology at the Molecular Level, Springer Verlag, for the Varna eneolithic cemetery, Bulgarian Black Sea Coast, Antiquity,
Berlin, 1993, pp. 1e37. in press.
[5] Y. Boyadziev, Chalcolithic stone architecture from Bulgaria, Archaeolog- [27] K.A. Hobson, R.T. Alisauskas, R.G. Clark, Stable-nitrogen isotope
ica Bulgarica VIII/1 (2004) 1e12. enrichment in avian tissues due to fasting and nutritional stress: implica-
[6] E. Bozilova, J. Atanassova, Paleoekologichni usloviya i istoriya na rasti- tions for isotopic analysis of diet, Condor 95 (1993) 388e394.
telnostta v raiona na Durankulashkoto ezero, in: H. Todorova (Ed.), [28] N.V. Honch, A palaeodietary and chronometric investigation of the
Durankulak Tom I, BAN, Sofia, 1989, pp. 197e204. prehistoric necropoleis of Varna and Durankulak, Bulgaria, unpublished
[7] E. Bozilova, S. Tonkov, Paleoecological evidence on the vegetation MSc dissertation, University of Oxford, 2004.
history and human occupation in the coastal area of Lake Durankulak, [29] N.V. Honch, T. Higham, J. Chapman, B. Gaydarska, H. Todorova, V.
Northeastern Bulgaria, in: H. Todorova (Ed.), Durankulak, Band II: Slavchev, Y. Yordanov, B. Dimitrova, Pontic chronologies and diets:
Die Prähistorischen Gräberfelder von Durankulak, Deutsches Archäolo- a scientific framework for understanding the Durankulak and Varna I
gisches Institut, Berlin, 2002, pp. 309e312. cemeteries, Bulgaria, in: L. Nikolova (Ed.), Black Sea Prehistory,
[8] C. Bronk Ramsey, T.F.G. Higham, A. Bowles, R.E.M. Hedges, Improve- in press.
ments to the pretreatment of bone at Oxford, Radiocarbon 46 (2004) [30] M.R. Howland, L.T. Corr, S.M.M. Young, V. Jones, S. Jim, N.J. van der
155e163. Merwe, Expression of the dietary isotope signal in the compound-specific
[9] A. Burke, V. Eisenmann, G.K. Ambler, The systematic position of Equus d13C values of pig bone lipids and amino acids, International Journal of
hydruntinus, an extinct species of Pleistocene equid, Quaternary Osteoarchaeology 13 (2003) 54e65.
Research 59 (2003) 459e469. [31] P. Iacumin, V. Nikolaev, L. Genoni, M. Ramigni, Y.G. Ryskovk,
[10] A. Cannon, H.P. Schwarcz, M. Knyf, Marine-based subsistence trends A. Longinelli, Stable isotope analyses of mammal skeletal remains of
and the stable isotope analysis of dog bones from Namu, British Colum- Holocene age from European Russia: a way to trace dietary and environ-
bia, Journal of Archaeological Science 26 (1999) 399e407. mental changes, Geobios 37 (2004) 37e47.
1504 N.V. Honch et al. / Journal of Archaeological Science 33 (2006) 1493e1504

[32] I. Ivanov, Varna and the birth of the European Civilization, Varna Camp cemetery site, Journal of Archaeological Science 25 (1998)
Necropolis: The Dawn of European Civilization, AGATÓ Publishers, 1247e1252.
Sofia, 2000, pp. 5e13. [47] M.P. Richards, J.A. Pearson, T.I. Molleson, N. Russell, L. Martin, Stable
[33] M.A. Katzenberg, A. Weber, Stable isotope ecology and palaeodiet in the isotope evidence of diet at Neolithic Çatalhöyük, Turkey, Journal of
Lake Baikal region of Siberia, Journal of Archaeological Science 26 Archaeological Science 30 (2003) 67e76.
(1999) 659. [48] H.P. Schwarcz, Some Theoretical aspects of isotope paleodiet studies,
[34] W.F. Libby, R. Berger, J. Mead, G. Alexander, J. Ross, Replacement rates Journal of Archaeological Science 18 (1991) 261e275.
for human tissue from atmospheric radiocarbon, Science 146 (1964) [49] H.P. Schwarcz, M.J. Schoeninger, Stable isotopic analyses in human
1170e1172. nutritional ecology, Yearbook of Physical Anthropology 34 (1991) 283e321.
[35] D. Lubell, M. Jackes, H. Schwarcz, M. Knyf, C. Meiklejohn, The Meso- [50] N. Spassov, N. Iliev, The animal bones from the prehistoric necropolis
lithic-Neolithic transition in Portugal: isotopic and dental evidence of near Durankulak (NE Bulgaria) and the latest record of Equus hydrunti-
diet, Journal of Archaeological Science 21 (1994) 201e216. nus regalia, in: H. Todorova (Ed.), Durankulak, Band II: Die Prähistori-
[36] E. Marinova, The new pollen core Lake Durankulak-3: a contribution to schen Gräberfelder von Durankulak, Deutsches Archäologisches Institut,
the vegetation history and human impact in Northeastern Bulgaria, in: Berlin, 2002, pp. 313e324.
S. Tonkov (Ed.), Aspects of Palynology and Palaeoecology: Festschrift [51] H. Todorova (Ed.), Durankulak, Band II: Die Prähistorischen Gräber-
in Honor of Elissaveta Bozilova, Pensoft Publishers, Sofia, 2003, pp. felder von Durankulak, Deutsches Archäologisches Institut, Berlin, 2002.
279e288. [52] H. Todorova, Tellsiedlung von Durankulak, Fritz Thiessen Stiftung
[37] M. Minagawa, E. Wada, Stepwise enrichment in 15N along food-chain: Jahresbericht, 1995/1996 (1997) 81-84.
further evidence and the relation between d15N and animal age, Geochi- [53] H. Todorova, Neolithic, Eneolithic, and transitional period in Bulgarian
mica et Cosmochimica Acta 48 (1984) 1135e1140. Prehistory, in: D.W. Bailey, I. Panayotov (Eds.), Prehistoric Bulgaria,
[38] M.L. Murray, M.J. Schoeninger, Diet, Status, and complex social struc- Prehistory Press, Madison, 1995.
ture in Iron Age central Europe: some contributions of bone chemistry, [54] H. Todorova, Ausgrabungen in Durankulak, Bezirk Tolbuchin
in: D.B. Gibson, M.N. Geselowitz (Eds.), Tribe and Polity in Late Prehis- (Bulgarien), in der Periode 1975e1981, Nachrichten aus Niedersachsens
toric Europe, Plenum Press, New York, 1988, pp. 155e176. Urgeschichte 52 (1983) 77e89.
[39] G. Nobis, Zur Fauna der frühneolitischen Siedlung Ovcharovo-gorata, [55] G.J. van Klinken, Bone collagen quality indicators for palaeodietary and
Bez. Târgovishte (NO-Bulgarien), Bonner Zoologische Beiträge 37 radiocarbon measurements, Journal of Archaeological Science 26 (1999)
(1986) 1e22. 687e695.
[40] K. Privat, Palaeoeconomy of the Eurasian Steppe: biomolecular studies, [56] G.J. van Klinken, M.P. Richards, R.E.M. Hedges, An overview of causes
unpublished DPhil dissertation, University of Oxford, 2004. for stable isotopic variations in past European human populations: environ-
[41] T. Prowse, H.P. Schwarcz, S. Saunders, R. Macchiarelli, L. Bondioli, Iso- mental, ecophysiological, and cultural effects, in: S.H. Ambrose,
topic paleodiet studies of skeletons from the Imperial Roman-age ceme- A. Katzenberg (Eds.), Biogeochemical Approaches to Palaeodietary Anal-
tery of Isola Sacra, Rome, Italy, Journal of Archaeological Science 31 ysis, Kluwer Academic/Plenum Publishers, New York, 2000, pp. 39e63.
(2004) 259e272. [57] G.J. van Klinken, H. van der Plicht, R.E.M. Hedges, Bond 13C/12C ratios
[42] A. Raduntcheva, La société dans les Balkans à l’âge du cuivre, Dossiers reflect (palaeo-)climatic variations, Geophysical Research Letters 21
Histoire et Archéologie 137 (1988) 46e55. (1994) 445e448.
[43] C. Renfrew, Varna and the emergence of wealth in prehistoric Europe, in: [58] I. Vörös, Wild equids from Early Holocene in the Carpathian Basin, Folia
A. Appadurai (Ed.), The Social Life of Things: Commodities in Cultural Archaeologia 32 (1981) 37e68.
Perspective, Cambridge University Press, Cambridge, 1986, pp. 141e168. [59] M. Yoneda, R. Suzuki, Y. Shibata, M. Morita, T. Sukegawa,
[44] C. Renfrew, Varna and the social context of early metallurgy, Antiquity N. Shigehara, T. Akazawa, Isotopic evidence of inland-water fishing by
52 (1978) 199e203. a Jomon population excavated from the Boji site, Nagano, Japan, Journal
[45] M.P. Richards, R.E.M. Hedges, Stable isotopic evidence for similarities of Archaeological Science 31 (2004) 97e107.
in the types of marine foods used by Late Mesolithic humans at sites [60] Y. Yordanov, B. Dimitrova, Results of an anthropological study of human
along the Atlantic Coast of Europe, Journal of Archaeological Science skeletal remains of the prehistoric necropolis in the vicinity of the village
26 (1999) 717e722. of Durankulak, in: H. Todorova (Ed.), Durankulak Band II: Die Prähis-
[46] M.P. Richards, R.E.M. Hedges, T.I. Molleson, J.C. Vogel, Stable iso- torischen Gräberfelder von Durankulak, Deutsches Archäologisches
tope analysis reveals variations in human diet at the Poundbury Institut, Berlin, 2002, pp. 325e347.

View publication stats