Journal of Archaeological Science 35 (2008) 2707–2714

Contents lists available at ScienceDirect

Journal of Archaeological Science

journal homepage: http://www.elsevier.com/locate/jas

Ancient DNA in human bones from Neolithic and Bronze Age sites in
Greece and Crete
Elizabeth R. Chilvers a, Abigail S. Bouwman a, Keri A. Brown a, Robert G. Arnott b,
A. John N.W. Prag c, Terence A. Brown a, *
a
Manchester Interdisciplinary Biocentre, Faculty of Life Sciences, University of Manchester, 131 Princess Street, Manchester, M1 7DN, UK
b
Centre for the History of Medicine, The Medical School, University of Birmingham, Birmingham, B15 2TT, UK
c
The Manchester Museum, University of Manchester, Oxford Road, Manchester, M13 9PL, UK

a r t i c l e i n f o a b s t r a c t

Article history: Attempts were made to detect ancient DNA (aDNA) in samples of 88 human skeletons from eight
Received 28 January 2008 Neolithic and Bronze Age sites in Greece and Crete. Ancient DNA was absent in specimens from Nea
Received in revised form 25 April 2008 Nikomedia, Lerna, Karaviádena (Zakro), Antron Grave Circle A and Mycenae Grave Circle A. For each of
Accepted 28 April 2008
three skeletons from Antron Grave Circle B that were sampled, polymerase chain reactions (PCRs) gave
products for nuclear but not mitochondrial DNA, but amplicon yield was low and inconsistent with
Keywords: replicate PCRs failing to give reproducible results. With specimens from Mycenae Grave Circle B,
Ancient DNA
evidence for mitochondrial aDNA was obtained for four of the 22 skeletons that were studied, and at
Bones
Bronze Age
Kouphovouno evidence for mitochondrial and/or nuclear aDNA was obtained with eight of the 20
DNA extraction skeletons that were examined. We conclude that, although aDNA might be present in some Eastern
Greece Mediterranean skeletons from later centuries of the Bronze Age, it is not commonly found in material
Mitochondrial DNA from this period and is likely to be absent from older material.
Polymerase chain reaction Ó 2008 Elsevier Ltd. All rights reserved.

1. Introduction Minoan and Mycenaean civilisations. Ancient DNA research could

The Neolithic and Bronze Ages of the Eastern Mediterranean
span approximately five millennia of great change during which
the early farming societies developed into the pre-Classical civili-
sations of Greece and the Aegean. The existence of human skeletal
remains from these periods raises the possibility that analysis of
ancient DNA (aDNA) could be used to address some of the relevant
archaeological questions. For example, it might be possible to use
genetic profiling to establish kinship relationships between groups
of burials (Brown, 2001) and, in conjunction with bone isotope
studies (Price, 1989), to identify individuals who might be incomers
to a particular site. This would be especially interesting if applied to
sites such as Grave Circles A and B at Mycenae, where the remains
are presumed to be of individuals who held high status in their
societies (Mylonas, 1957), and whose family relationships might
reveal the underlying processes by which such status was acquired.
Broader population studies might indicate affinities between the
people living in different areas at different periods, possibly
throwing light on questions such as the relationship between the
* Corresponding author: Tel.: þ44 161 306 4173; fax: þ44 161 306 5201.
E-mail address: terry.brown@manchester.ac.uk (T.A. Brown).
also answer long-standing questions about disease in the pre-
historic Aegean, in particular to test hypotheses regarding the
prevalence of malaria (Angel, 1966). As well as searching directly
for aDNA signatures of the malaria parasite in human bones
(Sallares and Gomzi, 2001), typing of globin gene mutations could
determine if the skeletal indications of anaemia seen in some
prehistoric Aegean skeletons are indeed due to genetic thalassae-
mia rather than the result of acquired iron deficiency anaemia
(Chilvers, 2004).
Progress in any of these areas is clearly dependent on the sur-
vival of aDNA in human skeletons from Eastern Mediterranean
sites. The most important consideration in this regard is not the
chronological age of the specimens but their thermal history, as
DNA degradation occurs more rapidly at higher temperatures. A
useful measure is ‘thermal age’ (Smith et al., 2003), which is cal-
culated from the temperature history of a site and its geographical
location. From experimental studies of DNA breakdown, together
with a consideration of the oldest authenticated detections of
aDNA, it has been estimated that the limit for DNA preservation is
approximately 19,000 years at 10  C: hence specimens from any
site that has a thermal age normalised to 10  C at >19,000 years are
unlikely to contain aDNA. When these calculations are applied to
the Eastern Mediterranean, a thermal age of 19,000 years at 10  C

0305-4403/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jas.2008.04.019
2708 E.R. Chilvers et al. / Journal of Archaeological Science 35 (2008) 2707–2714
corresponds to approximately 3600 chronological years (Chilvers,
2004). This makes it unlikely that aDNA will be present at Neolithic
sites, which predate 2000 BC, and suggests that the Bronze Age
sites of the Minoan palace period (c. 2000–1450 BC) and of Myce-
naean Greece (c. 1650–1150 BC) will be at the very limits for aDNA
preservation. At these sites we therefore anticipate that local fac-
tors such as exposure to water and time that has elapsed since
excavation (Pruvost et al., 2007) will be crucial in determining if
aDNA can be recovered.
Previous studies of bones and teeth from prehistoric Greece
support the prediction that aDNA is recoverable from some sites
dating to 2000 BC, with local factors having an important impact on
preservation (Evison et al., 1999; Brown et al., 2000; Evison, 2001).
Although carried out to the highest standards prevailing at the
time, these previous studies took place before stringent criteria for
aDNA authenticity had been established (Cooper and Poinar, 2000),
and it appears probable, in retrospect, that some of the ‘detections’
of aDNA described in these papers were in fact due to modern
contamination. Conversely, it is possible that some specimens that
gave negative results in these projects would have yielded au-
thentic aDNA had the highly-sensitive polymerase chain reactions
(PCRs) now available been used. We have therefore re-evaluated
the potential of aDNA research in the Eastern Mediterranean by
carrying out a systematic survey of aDNA preservation at eight
Neolithic and Bronze Age sites in Greece and Crete, examining 88
skeletons in total, and using the most up-to-date aDNA techniques.
2. Materials and methods
2.1. Sites and bone specimens
Bone specimens are listed in Table 1 and the locations of the
sites from which these were obtained are shown in Fig. 1. The early
Neolithic village of Nea Nikomedia is a multi-period settlement
mound located on the central Macedonian plain, close to the
south-west border of the Lake Ludias marshland and excavated in
1961–1964 (Wardle, 1996). Calibrated radiocarbon dates for the
earliest phases of occupation at the site fall into the range 6400–
6000 BC (Perles, 2001). In contrast with other early Neolithic sites,
burials have been found within the settlement, many in shallow,
irregular pits located outside the houses or in the rubble of older
collapsed houses rather than under the floors. Lerna is located in
the south-east Argolid, on the western shore of the Bay of Argos
and was occupied from the sixth to the first millennia BC; exca-
vations were carried out between 1952 and 1958. During the first
phases of occupation in the Neolithic, it is likely that the site was
some distance from the sea, but as sea levels continued to rise,
significant regions of the Argive plain were covered with water
and Lake Lerna was formed (Zangger, 1991), creating a marshy
environment in the vicinity of the site. Although the Neolithic
contexts yielded nine burials, the majority of burials date from the
Middle Bronze Age (2050/2000–1700/1675 BC). There are no
burials in the Early Bronze Age contexts and later burials are rare.
The Middle Bronze Age burials are usually found either next to the
houses or, in the case of some of the infant burials, under the floors
of houses. The great palace site of Zakro is located on the eastern
coast of Crete. Although the earliest evidence of occupation at the
site dates to Early Minoan III (2300–2100 BC), the palace which
dominates the region belongs to the Middle and Late Minoan
periods (Myers et al., 1992). During the construction of a road from
Epano Zakro to Kato Zakro a grave with what appears to be six
burials, dated to the Middle Minoan II period (1850–1700 BC), was
discovered at the site of Karaviádena, somewhat less than 1 km
south of the palace. It was subsequently excavated in 1994 and
may form part of a larger cemetery (R.G.A., unpublished). Antron
(Glypha Bay, Fthiotida) is located on the east coast of mainland
Greece. Grave Circles A and B are adjacent to one another, with
most of the burials in cist graves and dating to the Middle Helladic
III to Late Helladic IIA periods (1750–1450 BC) (excavated 1990–
1995; Papakonstantinou, 1999a,b). Mycenae is located in the
north-east of the Peloponnesos some 135 km from Athens. Grave
Circles A and B date to 1675–1500 BC, Grave Circle B predating A
with possibly a 50-year overlap between the two. The Grave
Circles therefore date to the very beginning of the Mycenaean age
at the boundary of the Middle to Late Helladic periods. Within
Grave Circle B, excavated in 1951, there is a development from
simple cist burials to larger, deeper and richer Shaft Graves, while
Grave Circle A, dug in 1876–1877, comprises six Shaft Graves
(Mylonas, 1957). Kouphovouno, located in Laconia just south of
Sparta, spans the Middle Neolithic to Late Bronze Age periods (c.
5000–1200 BC), and was the subject of a major excavation during
2001–2005. Twenty-seven burials were recovered, most of them
from a Middle Bronze Age cemetery (2000–1700 BC), mostly from
shallow earth graves (W. Cavanagh, personal communication).
2.2. DNA techniques
The outer 1–2 mm of each bone sample was removed with
a sterile scalpel (Bouwman et al., 2006) and the bone UV irradiated
(254 nm, 120,000 mJ cm2, 2  5 min with the bone rotated 180
between each exposure), to minimise the risk of contaminating
DNA on the surface being carried over to the extraction process.
Each bone was sealed in a DNA-free plastic bag and crushed into
powder. Between 0.49 and 0.51 g of bone powder was measured
into 1.5-ml microfuge tubes. Two or three additional empty tubes
were used as blanks for each set of extractions. DNA extractions
were carried out as previously described (Bouwman and Brown,
2005). All reagents and buffers were made up with ultrapure
water. A standard PCR of 50 ml contained 2.5 ml of bone extract, the
appropriate amount of MgCl2, 200 mM each dNTP, 100 ng each
primer, 1% bovine serum albumin, 1.25 units Taq polymerase (MBI
Fermentas) or AmpliTaq Gold DNA polymerase (Applied Bio-
systems), and 1  buffer (75 mM Tris–HCl pH 8.8, 20 mM
(NH4)2SO4, 0.01% Tween 20 for Taq; 150 mM Tris–HCl pH 8.0,
500 mM KCl for AmpliTaq Gold). PCRs designed to test for the
presence of inhibitory substances in bone extracts also contained
2.5 ml modern human DNA extract prepared from cheek cell sam-
ples using the method of Walsh et al. (1991). Cycling conditions for
all PCRs were: 4 min at 94  C; followed by 44 cycles of 1 min at the
appropriate annealing temperature, 1 min at 72  C and 1 min at
94  C; followed by 1 min at the annealing temperature and 10 min
at 72  C. Thirty-six PCRs were used in this study, targeting nine
mitochondrial DNA (mtDNA) loci spanning most of hypervariable
region 1 and part of region 2, 18 autosomal short tandem repeats
(STRs), five Y chromosome STRs, two Y chromosome single nu-
cleotide polymorphisms (SNPs), and two regions of the sex-specific
amelogenin locus. Primer sequences, MgCl2 concentrations and
annealing temperatures are given in Table 2. The Nea Nikomedia,
Lerna and Karaviádena specimens were tested with the mtDNA
PCRs MtD and MtH and the amelogenin PCR GA, individually using
Taq polymerase. The Antron, Mycenae and Kouphovouno bones
were tested with multiplex PCRs using AmpliTaq Gold DNA poly-
merase. Those PCRs marked in Table 2 with superscript were
combined in a triplex and a hexiplex system as previously pub-
lished (Butler et al., 2003; Coble and Butler, 2005). The Mgþþ
concentrations and annealing temperatures for all the other PCRs
were optimised by us and these PCRs combined in multiplex
systems that will be described elsewhere (A.S.B. and T.A.B.,
manuscript in preparation). PCRs were examined by electropho-
resis in 3% agarose or 11% polyacrylamide gels. Bands were purified
using Qiaquick columns (Qiagen), and DNA was ligated into pCR2.1
TOPO (Amersham) and cloned in Escherichia coli TOP10F0 cells.
 E.R. Chilvers et al. / Journal of Archaeological Science 35 (2008) 2707–2714 2709

Table 1
Bone specimens

Site Date Numbera Description

Nea Nikomedia 6400–6000 BC Infant 1 (A6/1) Long bone fragment
Infant 3 (U6/1A) Long bone fragment
Infant 5 (B6/2) Long bone fragment
NN4 (NN XLI D5/1 IV) Left clavicle fragment
NN13* (NN XIIIR7/1C) Short rib fragment
NN19 (NN XIX C9/1d) Long bone fragment
NN5 (NN L E3 V) Second metatarsel
NN23 (NN XXIII TX-5/1) Right second metatarsel
NN28 (XXVIII Ta2) Left third metatarsel
Lerna 6th millennium BC Ler 220 (EN) Mandible fragment
2050–1675 BC Ler 10 Cranial fragment
Ler 103 Long bone fragment
Ler 81 Radius fragment
Ler 125 Cranial fragment
Ler 48 Patella
Ler 203 Cranial fragment
Ler 72 Radius fragment
Karaviádena, 1850–1700 BC Burial 1a (A1-A2-K3, ZK15) Left talus
Zakro Burial 1b (A1-A2-K3, ZK16) Left talus
Burial 1c (A1-A2-K3, ZK17) Left talus
Burial 1d (A1-A2-K3, ZK18) Left talus
Burial 2 (A3, ZK8) Long bone fragment
Burial 3 (A4-A7, ZK10) Long bone fragment
Burial 4-5 (K1-A-A5, ZK19) Long bone fragment
Antron Grave 1750–1450 BC AXLVIIIbMH Skull fragment
Circle A AXII Long bone fragment
AI Long bone fragment
AXII Long bone fragment
AXLVIIIbLH Long bone fragment
ALIII Long bone fragment
Antron Grave 1750–1450 BC BII Long bone fragment
Circle B BIII Long bone fragment
BV Long bone fragment
Mycenae Grave 1600–1500 BC A3.1 Mandible
Circle A A3.2 Mandible
A3.3 Mandible
A4.1 Mandible
A4.2 Mandible
A4.3 Mandible
A4.22 Mandible
A4.27 Mandible
A5.25 Mandible
A5.26 Mandible
A6a Clavicle
A6b Clavicle
AM Unrecorded
Mycenae Grave 1675–1550 BC G51 Long bone fragment
Circle B B52 Long bone fragment
P53 Long bone fragment
H54 Long bone fragment
G55 Long bone fragment
L56 Long bone fragment
X57 Long bone fragment
G58 Long bone fragment
Z59 Long bone fragment
D60 Long bone fragment
D61 Long bone fragment
A62 Long bone fragment
Q63 Long bone fragment
N66 Long bone fragment
N66a Long bone fragment
I68 Long bone fragment
A69 Long bone fragment
K70 Long bone fragment
L70a Long bone fragment
L70a1 Long bone fragment
L70a2 Long bone fragment
L70a3 Long bone fragment
(continued on next page)
2710 E.R. Chilvers et al. / Journal of Archaeological Science 35 (2008) 2707–2714

Table 1 (continued)

Site Date Numbera Description

Kouphovouno 2000–1700 BC KE009 Tibia fragment
KE009B Tibia fragment
KE105 Humerus fragment
KE108 Radius fragment
KE171 Radius fragment
KE173 Humerus fragment
KE186 Tibia fragment
KE207 Tibia fragment
KE213L Humerus fragment
KE213U Fibula fragment
KE216 Rib fragment
KE220 Femur fragment
KE601 Femur fragment
KE704 Skull fragment
KE705 Humerus fragment
KE706 Ulna fragment
KE707 Humerus fragment
KE713 Tibia fragment
KE715 Tibia fragment
KE719 Tibia fragment
a
According to site, inventory or museum records. For osteology reports see: Nea Nikomedia, Angel (1973a), Lagia, (1993); Lerna, Angel (1971), Lagia (1993); Karaviádena,
Arnott and Morgan-Forster (in press); Antron, A. Papathanasiou (unpublished); Mycenae, Angel (1973b), Musgrave et al., 1995; Kouphovouno, A. Lagia (unpublished).

Recombinant plasmid DNA was purified using Qiaquick columns
and sequenced with the ABI Big Dye sequencing kit (Amersham).
2.3. Ancient DNA regime
We carried out the work in accordance with the standard cri-
teria of authenticity for aDNA research (Cooper and Poinar, 2000) as
far as was possible. Extractions and PCRs for the Nea Nikomedia,
Lerna and Karaviádena specimens were set up in laminar flow
cabinets in physically isolated labs; those for the Antron, Mycenae
and Kouphovouno specimens were set up in physically isolated labs
each with an ultrafiltered air supply maintaining positive dis-
placement pressure and a managed access system, with specimens
handled and extractions prepared within a Class II biological safety
cabinet and PCRs set up within a laminar flow hood. All extractions
were accompanied by negative controls in which the entire ex-
traction procedure was performed without bone material, and all
PCRs were accompanied by negative controls containing water
instead of DNA extract. The lengths of the template molecules in all
extracts giving positive PCRs were assessed by carrying out the MtZ
PCR (Table 2), which gives a product of 425 bp, longer than most
genuine aDNA molecules, even those from the best-preserved
material (O’Donoghue et al., 1996). The identity of every PCR
product was checked by sequencing of cloned amplicons. Because
of the small amounts of material that were available, it was not
possible to perform replicate extractions for all skeletons, nor was it
possible to divide the bone samples so that portions could be sent
to a second lab for independent testing, and, similarly, there was
insufficient material to carry out tests aimed at determining the
overall level of biomolecular preservation in the specimens. Cor-
roboration of the human results could not be sought through study
of associated animal remains, as no animal remains were available.
We did not quantify the number of starting templates in the ex-
tracts by real-time PCR, because knowing the number of starting
templates gives no assurance that these are aDNA rather than
contaminants (Malmström et al., 2005). We did, however, in-
troduce one additional precautionary step in that we removed the
outer 1–2 mm of each bone prior to preparation of extracts. We
have shown that even after extensive handling most of the con-
taminating DNA in a bone resides in the outer 1–2 mm (Bouwman
et al., 2006), and that very little redistribution occurs if the bone is
washed as in standard archaeological practice (M.M. Mundee, A.S.B.
and T.A.B., manuscript in preparation).

3. Results

The results are summarised in Table 3. No evidence of aDNA was

obtained with any of the specimens from Nea Nikomedia, Lerna,
Karaviádena, Antron Grave Circle A or Mycenae Grave Circle A. With
all but one specimen from these sites, PCRs failed to give a band of
the expected size. The exception was ZK8 from Karaviádena, one
extraction of which gave a band of the correct size with the MtD
PCR and an X-chromosome amplicon with the amelogenin PCR GA.
However, the sequence of the MtD amplicon was identical to that of
E.R.C. who studied this specimen, and ZK8 also gave the 425-bp
Fig. 1. Locations of sites. product with the MtZ PCR (again with the E.R.C. sequence). A
 E.R. Chilvers et al. / Journal of Archaeological Science 35 (2008) 2707–2714 2711

Table 2
Details of PCRs

Target Amplicon size (bp) Conditions Primers (50 –30 )

Mgþþ (mM) Annealing temp ( C)

Mitochondrial hypervariable region (locations in brackets)
MtA 119 2.0 54 CTAAGATTCTAATTTAAACTATTCTCTG
(15996–16114) GTGGCTGGCAGTAATGTACG
MtG 168 2.0 56 TTCATGGGGAAGCAGATTTGG
(16028–16195) ATGGGGAGGGGGTTTTGATGTGG
MtC 148 2.0 55 CCACCTGTAGTACATAAAAACCCA
(16147–16294) GTGGGTAGGTTTGTTGGTATCCTA
MtF 131 2.0 57 ACAGCAATCAACCCTCAACTATCA
(16210–16340) TGTGCTATGTACGGTAAATGGCTT
MtD 150 2.0 56 TAGGATACCAACAAACCTACCCAC
(16271–16420) TGATTTCACGGAGGATGGTG
MtH 131 2.0 55 CCATGCTTACAAGCAAGT
(16192–16322) TGGCTTTATGTACTATGTAC
MtW 117 2.5 55 ACCCTATTAACCACTCACGG
(16–132) GACAGATACTGCGACATAGG
MtV 137 2.0 55 ATAATAATAACAATTGAATGTCTGCAC
(232–368) TTCTTTGTTTTTGGGGTTTG
MtZ 425 2.0 55 CTAAGATTCTAATTTAAACTATTCTCTG
(15996–16420) TGATTTCACGGAGGATGGTG

Genomic STRs
CD4 86–121 2.0 56 CAAGCCTCTGTTGATTTCAT
ACCCAGATTGGACTGCAGT
D1S656 68–125 2.0 55 GTGTTGCTGAAGGGTCAACT
GAGAAATAGAATCACTAGGGAACC
a
D2S1338 94–154 2.0 55 TGGAAACAGAAATGGCTTGG
GATTGCAGGAGGGAAGGAAG
D3S1358 97–145 2.5 55 CAAGCCTCTGTTGATTTCAT
ACCCAGATTGGACTGCAGT
D5S818 133–169 2.0 55 GGTGATTTTCCTCTTTGGT
TGATTCCAATCATAGCCACA
D6S366 138–162 1.5 56 AGAGGTTACAGTGAGCCGAGATTG
GAAGTCCTAACAGAATGGAAGGTCC
D8S535 130–158 2.0 55 CATGTGACAAAAGCCACAC
AGACAGAAATATAGATGAGAATGCA
D8S1179 157–205 2.0 56 TTTTTGTATTTCATGTGTACATTCG
CGTAGCTATAATTAGTTCATTTTCA
D10S1248a 87–127 2.0 55 TTAATGAATTGAACAAATGAGTGAG
GCAACTCTGGTTGTATTGTCTTCAT
D10S2325 113–168 2.0 55 TCACGAAAGAAGCCTTCTG
GAGCTGAGAGATCACGCACT
a
D14S1434 73–101 2.0 55 TGTAATAACTCTACGACTGTCTGTCTG
GAATAGGAGGTGGATGGATGG
D16S539a 70–126 2.0 55 ATACAGACAGACAGACAGGTG
GCATGTATCTATCATCCATCTCT
D18S51a 101–197 2.0 55 TGAGTGACAAATTGAGACCTT
GTCTTACAATAACAGTTGCTACTATT
a
D22S1045 76–109 2.0 55 ATTTTCCCCGATGATAGTAGTCT
GCGAATGTATGATTGGCAATATTTTT
FGAa 140–290 2.0 55 AAATAAAATTAGGCATATTTACAAGC
GCTGAGTGATTTGTCTGTAATTG
TH01 A 146–190 2.0 55 GTGGGCTGAAAAGCTCCCGATTAT
GTGATTCCCATTGGCCTGTTCCTC
a
TH01 B 58–102 2.0 55 CCTGTTCCTCCCTTATTTCCC
GTTTCTTGGGAACACAGACTCCATGGTG
VWA 126–170 2.0 55 CCCTAGTGGATGATAAGAATAATCAGTATG
GGACAGATGATAAATACATAGGATGGATGG

Y STRs
DYS389 145–169 þ 261–293 2.0 54 CCAACTCTCATCTGTATTATCTAT
TTATCCCTGAGTAGTAGAAGAAT
DYS391 93–121 1.5 56 TTCAATCATACACCCATATCTGTC
GATAGAGGGATAGGTAGGCAGGC
DYS393 108–140 2.5 51 GTGGTCTTCTACTTGTGT
AACTCAAGTCCAAAAAAT
DYS426 92–98 2.5 57 CTCAAAGTATGAAAGCATGACCA
GGTGACAAGACGAGACTTTGTG
DYS460 101–121 2.5 55 GAGGAATCTGACACCTCTGACA
ATCCATATCATCTATCCTCTGCCTA

Y SNPs
M35 96 2.5 55 AGGGCATGGTCCCTTTCTAT
TCCATGCAGACTTTCGGAGT
M173 81 1.5 50 TTTTCTTACAATTCAAGGGCATTTAG
CTGAAAACAAAACACTGGCTTATCA

(continued on next page)

2712 E.R. Chilvers et al. / Journal of Archaeological Science 35 (2008) 2707–2714

Table 2 (continued)

Target Amplicon size (bp) Conditions Primers (50 –30 )

Mgþþ (mM) Annealing temp ( C)

Amelogenin
GA 106/112 2.0 55 CCTGGGCTCTGTAAAGAATAGT
ATCAGAGCTTAAACTGGGAAGCTG
MBa 126/132 2.0 55 CCCTGGGCTCTGTAAAGAATAGTG
ACACAGGCTTGAGGCCAAC
a
PCRs combined into multiplex systems described by Butler et al. (2003) and Coble and Butler (2005). All other PCRs except MtH were combined into multiplex PCRs
designed by us (A.S.B. and T.A.B., manuscript in preparation).

second extraction from this specimen yielded no PCR products.
These results suggest that the first extraction of ZK8 had become
contaminated with modern DNA. The possibility that aDNA was
present in bone extracts but masked by the presence of co-puri-
fying substances that were inhibitory to PCR was tested by ‘spiking’
PCRs of modern human DNA with bone extracts. Amplicon yield
with these control PCRs was unaffected by addition of any bone
extract, indicating that inhibitory substances were absent (Arnott
and Stuckey, 2003).
The results with the three specimens from Antron Grave Circle B
were inconsistent but could possibly indicate the presence of aDNA.
Although no mtDNA amplicons were obtained, two bands were
obtained when nuclear PCRs were carried out with extracts of BII
and BIII, and a range of products were obtained after nuclear PCRs
with BV. Replicate PCRs did not, however, give reproducible results
and in general the amplicon yields were weak, meaning that there
was insufficient material for the identities of these amplicons to be
checked by cloning and sequencing.
Nuclear amplification products were occasionally obtained with
specimens from Mycenae Grave Circle B, but too sporadically for
the results to be authenticated. With mitochondrial PCRs, 18 of the
22 samples never gave an amplification product of the correct size,
or if they did then that product was considered to be non-endog-
enous to the sample because it was accompanied by contaminated
negative controls, was entirely made up of sequences containing an
unusual mutation at position 16172 possessed by A.S.B., who per-
formed all these extractions and PCRs, or was not human mtDNA.
The other four samples (G55, G58, Z59 and A62) gave sequences
which were considered to derive, at least in part, from aDNA. These
results are described in detail elsewhere (Bouwman et al., in press).
The bones we studied from Kouphovouno were excavated
during 2001–2002 under conditions designed to minimise
contamination with modern DNA, and the excavator and A.S.B.
were the only people who handled these bones prior to transfer of
samples to the high-containment laboratory at Manchester. The
mtDNA haplotypes and nuclear STR genotypes of the excavator and
A.S.B. are known. It has been suggested that once all sequences
identical to those of individuals who have handled a specimen are
excluded, then any sequences that remain are likely to be genuine
aDNA (Sampietro et al., 2006). On this basis, preliminary results
indicate that at least two of the Kouphovouno specimens (KE009B
and KE105) contain mitochondrial aDNA, and six (KE009B, KE173,
KE207, KE601, KE706, KE715) contain nuclear aDNA. None of these
specimens gave a product with the 425-bp MtZ PCR. This work is
ongoing and the presence of nuclear DNA in some specimens that
have not yet yielded mtDNA amplicons is probably due to the
incompleteness of our current data rather than the preferential
survival of nuclear DNA.
4. Discussion
Validation of aDNA research has been discussed extensively in
the literature, with the ‘criteria of authenticity’ proposed by Poinar
and Cooper (2000) considered by many to be the gold standard
against which such work should be judged. These criteria were,
however, proposed by researchers more familiar with animal rather
than human aDNA, and they fail to take account of the realities of
biomolecular archaeology, in particular the problems posed by the
limited amount of material that is usually available for study. Mu-
seum curators are, understandably, unwilling to allow destructive
analysis of anything more than very small samples taken from
human specimens, and their reluctance is likely to become greater
with the growing debate regarding ‘ownership’ of human archae-
ological remains. The requirements within the ‘criteria of

Table 3
Summary of results

Site PCRs attempted with each specimena Evidence for aDNA

Nea Nikomedia MtD (2), MtH (2), GA (2) None
Lerna MtD (2), MtH (2), GA (2) None
Karaviádena MtD (2), MtH (2), GA (2) None
Antron Grave Circle A MtC (2), D2S11338 (2), D5S818, D10S1248, D14S1434, D16S539 (2), D18S51 (2), D22S1045, FGA (2), None
THO1 B (2), DYS426, M35, GA (2), MB (2)
Antron Grave Circle B MtC (2), D2S11338 (2), D5S818, D10S1248, D14S1434, D16S539 (2), D18S51 (2), D22S1045, FGA (2), No evidence
THO1 B (2), DYS426, M35, GA (2), MB (2) of mtDNA. Inconsistent
results for nuclear DNA in all three
bones studied
Mycenae Grave Circle A MtA (2), MtG (2), MtC(2), MtF, MtD (2), MtW, MtV, CD4, D1S656, D2S1338, D3S1358 (2), None
D5S818, D6S366, D8S535, D8S1179 (2), D10S1248, D10S2325 (2), D14S1434, D16S539, D18S51,
D22S1045, FGA, TH01 A (2), TH01 B, VWA (2), DYS389, DYS391, DYS393, DYS426 (2), DYS460 (2),
M35, M173 (2), GA (2), MB
Mycenae Grave Circle B MtA, MtG, MtC, MtF, MtD, MtW, MtV (2), CD4 (2), D1S656, D2S1338, D3S1358 (2), D5S818, D6S366, mtDNA in G55, G58, Z59 and A62
D8S535, D8S1179, D10S1248, D10S2325 (2), D14S1434, D16S539, D18S51, D22S1045, FGA, TH01 A, TH01 B,
VWA (2), DYS389, DYS391, DYS393, DYS426, DYS460 (2), M35, M173, GA, MB
Kouphovouno MtC (3), D2S1338, D10S1248, D14S1434, D16S539, D18S51, D22S1045, FGA, TH01 B, GA (3), MB mtDNA and/or nuclear DNA in 7 bones
a
Numbers in brackets indicate PCRs that were carried out more than once with each specimen.
 E.R. Chilvers et al. / Journal of Archaeological Science 35 (2008) 2707–2714
authenticity’ for multiple extractions and PCRs to check re-
producibility of results, replication of extractions and PCRs in
a second lab, and analysis of specimens to assess the overall degree
of biomolecular preservation, are reasonable if one is working with
kilograms of animal fossils but less easy to satisfy with a gram or so
of human bone. The rigorous demand that these requirements be
met when human remains are being studied is counter-productive
as it means that even when genuine aDNA has been recovered from
a human specimen it is impossible to ‘authenticate’ the detection.
Recognising this problem, Gilbert et al. (2005) have recommended
that biomolecular archaeologists take a cognitive and self-critical
approach to authentication of results, a proposal that we support
because, although it places the onus for authentication on the re-
searcher rather than on a set of externally-agreed criteria, it does
provide a framework within which archaeological research can
progress.
A key component of a cognitive approach to authentication of
aDNA detection is a consideration of the age and preservation
conditions of the specimens under study and the time that has
elapsed since their excavation, and an evaluation of whether these
factors make it possible for DNA to have survived. As temperature is
the primary determinant of the rate of DNA breakdown, the thermal
history (Smith et al., 2003) of a site can give an indication of the
likelihood of aDNA presence in specimens, but such analyses are
approximate at best due to difficulties in determining factors such
as seasonal temperature fluctuations and the precise conditions in
the microenvironment occupied by the buried specimens. However,
assessment of the thermal history of a site gives an indication of the
age beyond which specimens are unlikely to contain aDNA – placing
a large burden of proof on researchers claiming aDNA detections
with older material – and helps identify in which specimens aDNA
survival is possible, providing a starting point for self-criticism of
results. In this context, judgment of the authenticity of results at one
site is aided by information on the extent of DNA survival at other
sites within the same geographical region and hence likely to be of
similar thermal ages. In this paper, we provide this information for
Neolithic and Bronze Age sites in Greece and Crete, sites whose
thermal histories place them at the very limits of the expected
survival time for aDNA.
We found possible evidence for aDNA at three of the eight sites
that we studied. At Antron Grave Circle B we obtained nuclear but
not mitochondrial PCR products from extracts of each of the three
skeletons that we tested, but amplicon yield was low and in-
consistent with replicate PCRs failing to give reproducible results.
Because of the low amplicon yields it has proved difficult to obtain
clones for sequence analysis. In the absence of reproducible data we
are unable to provide any definite support for the notion that these
specimens contain aDNA and accept that the results may well be
due to contamination of the bones with modern DNA. If the results
do indicate the presence of aDNA in these specimens, then that
aDNA is very poorly preserved and unlikely to yield useful in-
formation. At Mycenae Grave Circle B, we obtained evidence for
mitochondrial aDNA in four of the 22 skeletons that we studied. We
believe that these results are genuine and present our cognitive and
self-critical assessment along with the archaeological implications
of the aDNA sequences elsewhere (Bouwman et al., in press). At
Kouphovouno we also obtained evidence for aDNA that, subject to
more detailed assessment, we currently believe to be genuine
because of our knowledge of the mitochondrial haplotypes and
nuclear genotypes of the only two individuals who could have
contaminated the bone samples.
Equally important are the negative results that we obtained. We
found no evidence whatsoever of aDNA in specimens from Nea
Nikomedia, Lerna, Karaviádena or Mycenae Grave Circle A. For Nea
Nikomedia and the Lerna 220 specimens this result is far from
surprising because, at 7000–8400 years in age, these bones are
2713
substantially older than the expected limit (3600 years) for aDNA
survival in Greece based on calculations of thermal age (Chilvers,
2004). The younger specimens from Lerna are dated to 2050–
1675 BC and hence closer to the 3600-year age limit, but the
marshy conditions that have prevailed in the vicinity of Lerna for at
least part of the period that these skeletons have been buried
suggests a relatively high moisture content which is likely to pro-
mote DNA degradation. The specimens from Karaviádena (1850–
1700 BC) and Mycenae Grave Circle A (1600–1500 BC) are also
close to the thermal age limit and hence possibly expected to show
some indication of aDNA survival, but those from Karaviádena
were poorly preserved at the time of excavation, being highly
fragmented, suggesting that overall biomolecular preservation
might be poor, and the Mycenae Grave Circle A bones, excavated in
1876–1877, have been housed in museums for 130 years, and it is
now clear that aDNA breakdown accelerates after excavation of
bones (Pruvost et al., 2007), suggesting that any aDNA present in
the Grave Circle A bones when Schliemann discovered them will
have degraded during the intervening decades.
It is, of course, easier to find explanations for the absence of
aDNA in specimens than it is to prove its presence, but we believe
that our failures to obtain any indications of aDNA in most of the
specimens that we studied provide an important message for bio-
molecular archaeologists studying skeletons from the Neolithic and
Bronze Age periods of the Eastern Mediterranean. We used opti-
mised PCR systems in order to maximise our chances of detecting
aDNA if it was present, but we also used a high-containment facility
and took scrupulous care to remove surface contamination from the
bone samples, to prevent cross-contamination with PCR products
from previous experiments, and to identify contamination that
remained. We also confirmed that negative results were not due to
inhibition of PCRs by substances co-purifying with aDNA. We con-
clude that, although aDNA might be present in some skeletons from
later centuries of the Eastern Mediterranean Bronze Age, it is not
commonly present in Eastern Mediterranean material from this
period and is likely to be absent from older material. We therefore
believe that all putative detections of aDNA from these periods of
Eastern Mediterranean prehistory require convincing authentica-
tion, whether through self-criticism of results or through adherence
to criteria of authenticity.
Acknowledgements
We thank K.A. Wardle (University of Birmingham) for bone
samples from Nea Nikomedia, M. Wiencke (American School of
Classical Studies, Athens) for material from Lerna, S. Chryssoulaki
(Ministry of Culture, Athens) for material from Karaviádena, Zakro,
M.Ph. Papakonstantinou (Ephorate of Prehistoric and Classical
Antiquities, Lamia) for samples from Antron, L. Papazoglou-
Manioudaki (National Archaeological Museum, Athens) for the
Mycenae Grave Circle A remains, E. Palaiologou and colleagues
(Ephorate of Prehistoric and Classical Antiquities, Nafplion) for
material from Mycenae Grave Circle B, and W.G. Cavanagh
(University of Nottingham), C.B. Mee (University of Liverpool),
A. Lagia and J. Renard (Université de Montpelier) for bones from
Kouphovouno. We also thank M.J. Collins (York) for help with cal-
culation of thermal ages. The work was funded by the Institute for
Aegean Prehistory, the Wellcome Trust and the Leverhulme Trust.
References
Angel, J.L., 1966. Porotic hyperostosis, anaemias, malarias, and marshes in the
prehistoric Eastern Mediterranean. Science 153, 760–763.
Angel, J.L., 1971. Lerna. Vol. II: The People. American School of Classical Studies at
Athens, Princeton.
2714 E.R. Chilvers et al. / Journal of Archaeological Science 35 (2008) 2707–2714
Angel, J.L., 1973. Early Neolithic people of Nea Nikomedia. In: Schwidetzky, I. (Ed.),
Die Anfänge des Neolithikums vom Orient bis Nord Europa. Böhlau Verlag,
Cologne, pp. 103–112.
Angel, J.L., 1973. Human skeletons from grave circles at Mycenae. In: Mylonas, G.E.
(Ed.), Ho Taphikos Kyklos B ton Mykenon, vol. I. Athens Archaeological Society,
Athens, pp. 379–397.
Arnott, R., Morgan-Forster, A., in press. Malaria and the Prehistory of the Aegean
(BAR International Series). Archaeopress, Oxford.
Arnott, R., Stuckey, E., 2003. Finding the Hydra: the search for malaria in the
prehistoric Aegean. In: Foster, K., Laffineur, R. (Eds.), Metron: Measuring the
Aegean Bronze Age. Aegeum, 24. University of Liège, Liège, pp. 209–215.
Bouwman, A.S., Brown, T.A., 2005. The limits of biomolecular palaeopathology:
ancient DNA cannot be used to study venereal syphilis. J. Archaeol. Sci. 32, 691–
702.
Bouwman, A.S., Chilvers, E.R., Brown, K.A., Brown, T.A., 2006. Identification of the
authentic ancient DNA sequence in a human bone contaminated with modern
DNA. Am. J. Phys. Anthropol 131, 428–431.
Bouwman, A.S., Brown, K.A., Prag, A.J.N.W., Brown, T.A. Kinship between burials
from Grave Circle B at Mycenae revealed by ancient DNA typing. J. Archaeol. Sci.,
in press, doi: 10.1016/j.jas.2008.04.010.
Brown, T.A., 2001. Ancient DNA. In: Brothwell, D.R., Pollard, A.M. (Eds.), Handbook
of Archaeological Sciences. Wiley, Chichester, pp. 301–311.
Brown, T.A., Brown, K.A., Flaherty, C.F., Little, L.M., Prag, A.J.N.W., 2000. DNA analysis
of bones from Grave Circle B at Mycenae: a first report. Annu. Br. Sch. Athens 95,
115–119.
Butler, J.M., Chen, Y., McCord, B.R., 2003. The development of reduced size STR
amplicons as tools for analysis of degraded DNA. J. Forensic Sci. 48, 1054–
1064.
Chilvers, E.R., 2004. Ancient DNA and palaeopathology: malaria and Ancient Greece.
PhD thesis, University of Manchester Institute of Science and Technology.
Coble, M.D., Butler, J.M., 2005. Characterization of new miniSTR loci to aid analysis
of degraded DNA. J. Forensic Sci. 50, 43–53.
Cooper, A., Poinar, H., 2000. Ancient DNA: do it right or not at all. Science 289, 1139.
Evison, M.P., 2001. Ancient DNA in Greece: problems and prospects. J. Radioanal.
Nucl. Chem. 247, 673–678.
Evison, M.P., Fieller, N.J., Smillie, D.M., 1999. Ancient HLA: a preliminary survey. Anc.
Biomol. 3, 1–28.
Gilbert, M.T.P., Bandelt, H.-J., Hofreiter, M., Barnes, I., 2005. Assessing ancient DNA
studies. Trends Ecol. Evol. 20, 541–544.
Lagia, A., 1993. Differential diagnosis of the three main types of anaemia (thal-
assaemia, sickle cell anaemia, iron deficiency anaemia) based on macroscopic
and radiographic skeletal characteristics. MSc dissertation, University of
Bradford.
Malmström, H., Storå, J., Dalén, L., Holmlund, G., Götherström, A., 2005. Extensive
human DNA contamination of extracts from ancient dog bones and teeth. Mol.
Biol. Evol. 22, 2040–2047.
Musgrave, J.H., Neave, R.A.H., Prag, A.J.N.W., 1995. Seven faces from Grave Circle B at
Mycenae. Annu. Br. Sch. Athens 90, 107–136.
Myers, J.W., Myers, E.E., Cadogan, G. (Eds.), 1992. The Aerial Atlas of Ancient Crete.
Thames and Hudson, London.
Mylonas, G.E., 1957. Ancient Mycenae: The Capital City of Agamemnon. Routledge &
Kegan Paul, London.
O’Donoghue, K., Clapham, A., Evershed, R.P., Brown, T.A., 1996. Remarkable pres-
ervation of biomolecules in ancient radish seeds. Proc. R. Soc. Lond. B Biol. Sci.
263, 541–547.
Papakonstantinou, M. Ph., 1999. The Grave Circle B of Antron: preliminary report.
In: Betancourt, P.P., Karageorghis, V., Laffineur, R., Niemeyer, W.-D. (Eds.), Me-
letemata, Studies in Aegean Archaeology Presented to Malcolm H. Wiener as he
Enters his 65th year. Aegaeum, 20. University of Liège, Liège, pp. 626–631.
Papakonstantinou, M.Ph., 1999b. Ho Taphikos Kyklos A tis Antronas. Pepragmena
tou A Diepistemonikou Symposiou, He Periphereia tou Mykenaikou Kosmou,
(Lamia 1994), Lamia, pp. 171–180.
Perles, C., 2001. The Early Neolithic in Greece. Cambridge University Press,
Cambridge.
Price, T.D. (Ed.), 1989. The Chemistry of Prehistoric Bone. Cambridge University
Press, New York.
Pruvost, M., Schwarz, R., Bessa Correia, V., Champlot, S., Braguier, S., Morel, N.,
Fernandez-Jalvo, Y., Grange, T., Geigl, E.-M., 2007. Freshly excavated fossil bones
are best for amplification of ancient DNA. Proc. Natl. Acad. Sci. USA 104, 739–
744.
Sallares, R., Gomzi, S., 2001. Biomolecular archaeology of malaria. Anc. Biomol. 3,
195–213.
Sampietro, M.L., Gilbert, M.T.P., Lao, O., Caramelli, D., Lari, M., Bertranpetit, J., La-
lueza-Fox, C., 2006. Tracking down human contamination in ancient human
teeth. Mol. Biol. Evol. 23, 1801–1807.
Smith, C.I., Chamberlain, A.T., Riley, M.S., Stringer, C., Collins, M.J., 2003. The thermal
history of human fossils and the likelihood of successful DNA amplification. J.
Hum. Evol. 45, 203–217.
Walsh, P.S., Metzger, D.A., Higuchi, R., 1991. Chelex 100 as a medium for simple
extraction of DNA for PCR-based typing from forensic material. Biotechniques
10, 506–513.
Wardle, K.A. (Ed.), 1996. Nea Nikomedia 1: The Excavation of an Early Neolithic
Village in Northern Greece 1961–1964. BSA Supplementary Volume 25. British
School at Athens, Athens.
Zangger, E., 1991. Prehistoric coastal environments in Greece: the vanished
landscapes of Dimini Bay and Lake Lerna. J. Field Archaeol. 18, 1–15.