Τesting models for the beginnings of the Aurignacian and the advent of figurative art and music: the radiocarbon chronology of Geißenklösterle.

Abstract

The German site of Geißenklösterle is crucial to debates concerning the European Middle to Upper Palaeolithic transition and the origins of the Aurignacian in Europe. Previous dates from the site are central to an important hypothesis, the Kulturpumpe model, which posits that the Swabian Jura was an area where crucial behavioural developments took place and then spread to other parts of Europe. The previous chronology (critical to the model), is based mainly on radiocarbon dating, but remains poorly constrained due to the dating resolution and the variability of dates. The cause of these problems is disputed, but two principal explanations have been proposed: a) larger than expected variations in the production of atmospheric radiocarbon, and b) taphonomic influences in the site mixing the bones that were dated into different parts of the site. We reinvestigate the chronology using a new series of radiocarbon determinations obtained from the Mousterian, Aurignacian and Gravettian levels. The results strongly imply that the previous dates were affected by insufficient decontamination of the bone collagen prior to dating. Using an ultrafiltration protocol the chronometric picture becomes much clearer. Comparison of the results against other recently dated sites in other parts of Europe suggests the Early Aurignacian levels are earlier than other sites in the south of France and Italy, but not as early as recently dated sites which suggest a pre-Aurignacian dispersal of modern humans to Italy by ∼45000 cal BP. They are consistent with the importance of the Danube Corridor as a key route for the movement of people and ideas. The new dates fail to refute the Kulturpumpe model and suggest that Swabian Jura is a region that contributed significantly to the evolution of symbolic behaviour as indicated by early evidence for figurative art, music and mythical imagery.

Full text

Page 1

Τesting models for the beginnings of the Aurignacian and the advent of figurative
art and music: The radiocarbon chronology of Geißenklösterle
Thomas Highama,*, Laura Basellb, Roger Jacobic,d,1, Rachel Wooda,e, Christopher Bronk Ramseya,
Nicholas J. Conardf,g,*
aOxford Radiocarbon Accelerator Unit, Research Laboratory for Archaeology and the History of Art, University of Oxford, Oxford OX1 3QY, UK
bPalaeoenvironmental Laboratory at the University of Southampton (PLUS), Building 44, University of Southampton, University Road, Southampton SO17 1BJ, UK
cThe Natural History Museum, Cromwell Road, London SW7 5BD, UK
dDepartment of Prehistory and Europe (Quaternary Section), The British Museum London N1 5QJ, UK
eResearch School for Earth Sciences, Australian National University, Canberra, Australia
fAbt. Ältere Urgeschichte und Quartärökologie, Universität Tübingen, Schloss Hohentübingen, 72070 Tübingen, Germany
gTübingen Senckenberg Center for Human Evolution and Paleoecology, Schloss Hohentübingen, 72070 Tübingen, Germany
a r t i c l e i n f o
Article history:
Received 12 May 2011
Accepted 14 March 2012
Available online
Keywords:
AMS radiocarbon dating
Middle and upper Palaeolithic
Pre-treatment chemistry
Ultrafiltration
Bone collagen
Swabian Jura
a b s t r a c t
The German site of Geißenklösterle is crucial to debates concerning the European Middle to Upper
Palaeolithic transition and the origins of the Aurignacian in Europe. Previous dates from the site are
central to an important hypothesis, the Kulturpumpe model, which posits that the Swabian Jura was an
area where crucial behavioural developments took place and then spread to other parts of Europe. The
previous chronology (critical to the model), is based mainly on radiocarbon dating, but remains poorly
constrained due to the dating resolution and the variability of dates. The cause of these problems is
disputed, but two principal explanations have been proposed: a) larger than expected variations in the
production of atmospheric radiocarbon, and b) taphonomic influences in the site mixing the bones that
were dated into different parts of the site. We reinvestigate the chronology using a new series of
radiocarbon determinations obtained from the Mousterian, Aurignacian and Gravettian levels. The
results strongly imply that the previous dates were affected by insufficient decontamination of the bone
collagen prior to dating. Using an ultrafiltration protocol the chronometric picture becomes much clearer.
Comparison of the results against other recently dated sites in other parts of Europe suggests the Early
Aurignacian levels are earlier than other sites in the south of France and Italy, but not as early as recently
dated sites which suggest a pre-Aurignacian dispersal of modern humans to Italy by w45000 cal BP. They
are consistent with the importance of the Danube Corridor as a key route for the movement of people
and ideas. The new dates fail to refute the Kulturpumpe model and suggest that Swabian Jura is a region
that contributed significantly to the evolution of symbolic behaviour as indicated by early evidence for
figurative art, music and mythical imagery.
? 2012 Elsevier Ltd. All rights reserved.
Introduction
Debate surrounds the nature and timing of the transition from
the Middle to the Upper Palaeolithic in Europe. Most scholars
accept that the transformation from a Neanderthal-dominated
Europe to one exclusively peopled by anatomically modern
humans (AMH) occurred between w30e45,000 cal BP (calibrated
years before present) (Mellars, 1999; Jöris and Street, 2008). Key
areas of contention focus instead on the transition process, the
routes by which AMH expanded across Europe, and the spatial and
temporal development of the Aurignacian. Debate also continues
on the extent of cultural and genetic interaction between Nean-
derthals and AMH and how Neanderthals went extinct (Conard,
2006). Closely related to these issues is the challenging question
of behavioural modernity, a term linked solely to AMH by some
(e.g., Mellars, 1999, 2005) and to both groups by others (e.g.,
D’Errico et al., 1998; Zilhão and D’Errico, 1999; D’Errico, 2003;
Zilhão, 2006; Langley et al., 2008). More specifically, the early
radiocarbonand thermoluminescence
Geißenklösterle have played a key role in developing the Danube
(TL)datesfrom
* Corresponding authors.
E-mail addresses: thomas.higham@rlaha.ox.ac.uk (T. Higham), nicholas.conard@
uni-tuebingen.de (N.J. Conard).
1Deceased.
Contents lists available at SciVerse ScienceDirect
Journal of Human Evolution
journal homepage: www.elsevier.com/locate/jhevol
0047-2484/$ e see front matter ? 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jhevol.2012.03.003
Journal of Human Evolution xxx (2012) 1e13
Please cite this article in press as: Higham, T., et al., Τesting models for the beginnings of the Aurignacian and the advent of figurative art and
music: The radiocarbon chronology of Geißenklösterle, Journal of Human Evolution (2012), doi:10.1016/j.jhevol.2012.03.003

Page 2

Corridor and Kulturpumpe hypotheses (Richter et al., 2000; Conard
and Bolus, 2003). The first model hypothesises that the Danube
Valley served as a main artery for migrations into Central Europe,
while the second views the Swabian region as making important
contributions to the evolution of complex symbolic behaviour, as
indicated by the early presence of figurative art, musical instru-
ments, mythical imagery and three-dimensionally shaped personal
ornaments (Conard and Bolus, 2006; Conard, 2009; Conard et al.,
2009). Testing these hypotheses is only possible via a reliable
high-resolution chronology for the wider region.
Unfortunately, it is now apparent that the radiocarbon record,
constructed over the last 60 years, is significantly flawed and
inadequate for rigorously testing these models. This is due to the
combined effects of incomplete removal of contamination and
the difficulties encountered when dating samples very close to the
measurement limit (Higham, 2011). This was either not recognised,
or not adequately addressed, at the time of dating. In addition,
many of the determinations available for the Middle to Upper
Palaeolithic are often only useful in the broadest chronological
sense because of measurement imprecision. The development of
more refined methodological approaches has had a significant
effect in improving accuracy. The application of ultrafiltration for
dating bone, and ABOx-SC methods for dating charcoal have shown
for some sites, even those recently dated, that a large proportion of
dates may be aberrant (Jacobi et al., 2006; Higham et al., 2006a,
2009; Brock and Higham, 2009; Douka et al., 2010; Higham,
2011; Wood et al., in press).
Reliability of previous dates from Geißenklösterle:
stratigraphy, taphonomy and the ‘Middle Palaeolithic dating
anomaly’
The Swabian Jura of Germany is of particular interest because of
a concentration of Middle to Upper Palaeolithic sites including
Geißenklösterle, Hohle Fels, Vogelherd and Hohlenstein-Stadel.
Geißenklösterle lies in the former Danube Valley through which
the Ach River flows today (see Fig.1 for sites mentioned in the text).
The cave is situated high above the river on the southern side of the
valley near Blaubeuren, in the Swabian Jura of southwestern
Germany. The site was initially excavated by Eberhard Wagner in
1973. Joachim Hahn (1988) continued work at the site over 14
seasons of excavation between 1974 and 1991. Conard and
colleagues renewed fieldwork at the site from 2000 to 2002
(Conard and Malina, 2003).
Geißenklösterle comprises a sequence of archaeological levels
spanning the Middle Palaeolithic to the Magdalenian (Fig. 2). The
sitewas divided into a series of 19 geological horizons (GH) and five
archaeological horizons (AH, numbered I to V), which lie within
them. Within the AH are further sub-units. The key Early Auri-
gnacian layers focussed on here are within AH III. Between AH IV,
the uppermost Mousterian horizon, and AH III, there is a culturally
nearlysterile horizon. AH II comprises the Upper Aurignacian. Hahn
(1988) originally attributed the AH III lithic corpus to the Proto-
Aurignacian, but it is clear now that this is not the case. The lithic
evidence comprises numerous carinated and nosed end scrapers,
burins, and a small number of worked bone, ivory and antler arti-
facts (Hahn, 1988) typical of the Early Aurignacian (Conard and
Bolus, 2006; Teyssandier et al., 2006). Although several personal
ornaments were excavated, AH III lacked the range of bone and
antler ornaments, artworks and flutes found in AH II. This being
said, evidence for ivory working is more common in AH III than in
AH II (Bolus and Conard, 2001). Although the analysis of ivory from
water-screened samples has not yet been completed, the current
data shows that the lower Aurignacian contains 1015 pieces of ivory
weighing 2.42 kg, while the upper Aurignacian contains 478 pieces
of ivory weighing 0.78 kg. Both the absolute numbers and
proportions of workedivory pieces arehigher in AH III than in AH II,
indicating that the intensity of ivory working was much higher
during the formation of AH III (Conard et al., 2003; Münzel, in
press). During the most recent phase of excavation, the presence
Figure 1. Location of the site of Geißenklösterle, Germany, and other sites mentioned in the text.
T. Higham et al. / Journal of Human Evolution xxx (2012) 1e13
2
Please cite this article in press as: Higham, T., et al., Τesting models for the beginnings of the Aurignacian and the advent of figurative art and
music: The radiocarbon chronology of Geißenklösterle, Journal of Human Evolution (2012), doi:10.1016/j.jhevol.2012.03.003

Page 3

of ivory working debris at the base of the Aurignacian sequencewas
repeatedly noted, which indicates that the manufacture of ivory
artifacts was part of the cultural repertoire of the local Aurignacian
people from the start.
Conard and Bolus (2003, 2008) have published more than 80
radiocarbon determinations of bone from several radiocarbon
laboratories in an attempt to build a coherent chronometric
sequence and to test the integrity of the stratigraphic sequence at
the site. Some of the initial radiocarbon results from the Aurigna-
cian horizons of the site were very early, between w36e40 ka
(thousands of years ago) BP. Electron spin resonance (ESR) and TL
dates obtained by Richter et al. (2000) also supported a very early
date for some of these levels. These ages have formed the basis of
the Danube Corridor and Kulturpumpe hypothesis. As mentioned
above, these models suggest that the Danube River served as an
arterial route for facilitating the movement of early AMH into
Central Europe from around w40 ka BP and that important cultural
innovations of the Aurignacian developed in the Swabian Jura and
Upper Danube region (Conard, 2002; Conard and Bolus, 2003).
In Fig. 3, all of the radiocarbon results obtained thus far from the
site are shown in calendar years, based on the new INTCAL09 curve
(data from Conard and Bolus, 2008; calibration curve after Reimer
et al., 2009). They show wide variation. Several appear inconsistent
when considered in terms of the stratigraphic sequence of the site.
The reasons for this have remained elusive. Conard and Bolus
(2003) suggested that it might be caused by significant variations
in the production rate of radiocarbon, as attested strongly at the
time in the datasets obtained by Beck et al. (2001) and Voelker et al.
(2000). They coined the phrase ‘the Middle Palaeolithic Dating
Anomaly’ to explain the tendency for the variation in atmospheric
radiocarbon concentration during this period to produce variable
results in archaeological dates (Conard and Bolus, 2003). Others
have interpreted the evidence differently. Zilhão and D’Errico
(1999, 2003a, b), for example, consider the most important influ-
ence at the site to be post-depositional mixing and taphonomy,
caused by the cryoturbation of sediments and the influence of
periodic rock falls. Hahn (1988) also recognised this as an important
factor, and always stressed the difficulties in separating micro-
stratigraphic levels within the excavation sequence. Zilhão and
D’Errico (1999, 2003a, b) interpreted the radiocarbon determina-
tions literally, arguing that they indicated a slow rate of sediment
deposition and that repeated human occupations could have
resulted in intermixing of animal and human deposits. They
concluded that AH II and AH III represent several different occu-
pation events, resulting in the scatter of radiocarbon results. In
their view, the radiocarbon dates are accurate, but their original
context is blurred, so that, “no simple explanation exists for these
results and it is clear that none will be found unless the vertical
displacement of items is duly accounted for as part of the problem”
(Zilhão and D’Errico, 2003a: 333).
Meticulous piece-plotting of more than 22,000 excavated items
was undertaken at the site, as well as refitting studies of lithic
artefacts (Hahn, 1988; Conard et al., 2003). Based on this evidence,
Hahn (1988) estimated that about 60% of the excavated material
had remained in situ, that w40% of material had been moved from
its original context, but only 7% appeared to have moved between
AH II and AH III. Hahn, however, also stressed that some apparent
mixing is the result of the difficulty in defining meaningful strati-
graphic units within deposits that contained an abundance of
coarse limestone rubble from the weathering of the walls and
ceiling of the cave. Hahn described the importance of what he
called excavation error in explaining part of the mixing between
the excavated strata. The most recent phase of excavation under
Conard’s direction repeatedly documented that the fine stratig-
raphy of Hahn’s excavation was not reliable, confirming that Hahn
Figure 2. Stratigraphic schema for Geißenklösterle, showing the archaeological
horizons (AH) at the site. AH III and II correspond with Early Aurignacian levels,
whilst AH IV is a Mousterian horizon. Above AH II is a series of superposed Gravettian
layers.
T. Higham et al. / Journal of Human Evolution xxx (2012) 1e13
3
Please cite this article in press as: Higham, T., et al., Τesting models for the beginnings of the Aurignacian and the advent of figurative art and
music: The radiocarbon chronology of Geißenklösterle, Journal of Human Evolution (2012), doi:10.1016/j.jhevol.2012.03.003

Page 4

was justified in combining the many sub-units of the Aurignacian
to form two major units, AH II and AH III (Hahn, 1988). Hahn was
acutely aware of these issues and never suggested that the sub-
units of the Aurignacian formed closed assemblages or reliable
cultural stratigraphic units. Thus, the refits between the strati-
graphic units only served to confirm the difficulty of executing
a high-resolution excavation in these kinds of sediments.
Plots of the many piece-plotted finds from refitted reduction
sequences, however, show that the artefacts have a high degree of
stratigraphic integrity (Conard et al., 2003). Further research by
Teyssandier and Liolios (2003) and Teyssandier et al. (2006) also
considered the chaîne opératoire of the Aurignacian lithic corpus
from the site and its implications. They showed that for AH III, there
is a coherent lithic reduction strategy that suggests a low likelihood
of contamination and movement derived from the Aurignacian
horizonabove. They have also shown the lithics fromboth AH II and
III to be unequivocally Early Aurignacian in character, consisting of
many carinated/nosed pieces with blade debitage and comparable
with material from similar industries in Aquitaine. Their work
suggests that cultural material in AH III originating from AH II is
scarce. In addition, the results of micro-morphological analyses
indicate that little mixing has occurred between the find horizons
and that no mixing has taken place between objects from the
Aurignacian and the underlying Mousterian (Conard et al., 2003;
Goldberg and Conard, in press). Taken together, these studies
suggest that the archaeological material from the Aurignacian of
Geißenklösterle has not been strongly altered by taphonomic
processes (Hahn, 1988). Whilst acknowledging that natural and
cultural transformations of the distribution of the objects in the
Aurignacian horizons took place, and the excavations of these
deposits were by their very nature imperfect, the degree of mixing
was not substantial and the dating results cannot be interpreted as
being primarily the result of major reworking of the deposits. This
conclusion is important because it requires us to look for other
explanations for the radiocarbon record at Geißenklösterle.
More recent work has also shed light on the ‘Middle Palaeolithic
Dating Anomaly’ hypothesis (Conard and Bolus, 2003). Variation in
atmospheric14C production is widely attested in younger parts of
the radiocarbon timescale, where it is corrected for by the use of
calibration curves built on precise dendrochronologically-dated
oak and pine back to w12,600 BP (Reimer et al., 2009). Prior to
this, reliable calibration has proven elusive, with some records
showing great variability. At Geißenklösterle, the variability in the
radiocarbon dates was attributed in part to particularly large
discrepancies in14C production as attested in the data published by
Beck et al. (2001), and the14C dataset from their Bahamian spe-
leothem GB89-24-1. However, recent research has shown that it
was not possible to reproduce the results (Hoffmann et al., 2010)
and that the Beck et al. (2001) dataset was affected by an incorrect
radiocarbon blank value being subtracted from the dated samples.
A new and revised speleothem record shows fewer dramatic
oscillations (Hoffmann et al., 2010), although there are still parts
that do show variability. This record has not been included in the
latest INTCAL09 calibration dataset (see below), but parts of it may
be included in subsequent iterations (Reimer et al., 2009). More
recently obtained marine-derived datasets such as thatof Fairbanks
et al. (2005) obtained from coral from Barbados and the Cariaco
basin record of Hughen et al. (2004, 2006) show none of the wide
variations of the type shown by Beck et al. (2001), although the
fluctuations in their14C records may be dampened compared with
equivalent terrestrial archives. Recent reanalysis of the key Lake
Suigetsu terrestrial lake sediment sequence in Japan suggests that
the agreement between INTCAL09 and the Suigetsu sequence is
acceptably good when the 1983 Suigetsu dataset is statistically
fitted to INTCAL09 (Staff et al., 2010). The new full dataset from
Figure 3. Calibrated radiocarbon determinations from Geißenklösterle published by
Conard and Bolus (2008), prior to our work. Calibrated results are divided by cultural
association, as outlined in Conard and Bolus (2003: Table 1).
T. Higham et al. / Journal of Human Evolution xxx (2012) 1e13
4
Please cite this article in press as: Higham, T., et al., Τesting models for the beginnings of the Aurignacian and the advent of figurative art and
music: The radiocarbon chronology of Geißenklösterle, Journal of Human Evolution (2012), doi:10.1016/j.jhevol.2012.03.003

Page 5

Suigetsu is yet to be published but it appears that for the most part
it does not yield the substantial
through the period of interest, at least of the size previously seen.
Taken together, the datasets used to produce the current interim
curve do not vary by the order of magnitude previously identified
by Beck et al. (2001). Further work is clearly required, and is
ongoing. A hint that there might be more significant deviations is
provided by Hajdas et al. (2012) based on their data obtained from
a core in the Mediterranean Sea (see also Giaccio et al., 2006).
Significant variation in14C production occurs around the time that
tephra from the Campanian Ignimbrite appears in the core.
Modelling shows that the amplitudes of these variations cannot be
reproduced using box diffusion models, however. In addition, as we
show below, the Middle to Upper Palaeolithic transitional sequence
at Geißenklösterle pre-dates the date of the eruption by several
millennia. Although undoubtedly an important factor, then, radio-
carbon production variability would not appear to be the principal
reason in explaining the spread in the initial radiocarbon results
from Geißenklösterle shown in Fig. 3.
14C production rate variations
Materials and methods
It is now widely known that the radiocarbon method has been
plagued with problems in its proper application over the period of
the Middle to Upper Palaeolithic transition, because the low
amount of residual14C (<3%, equivalent to w30,000 BP2) makes the
effect of young contaminating carbon significant. The Oxford
Radiocarbon Accelerator Unit (ORAU), and other laboratories, have
workedfor many years toimprove the datingof material of this age,
including those from European Palaeolithic sites. In 2001, ORAU
adopted an ultrafiltration technique to improve the routine
collagen purification in our facility (based on Brown et al.,1988; see
also; Bronk Ramsey et al., 2004a; Higham et al., 2006b). This
method allows the separation of lower molecular weight compo-
nents from the gelatin extracted from archaeological bones using
the Longin collagen method (Longin, 1971). The application of the
method to the British Middle to Upper Palaeolithic has been
significant (Jacobi et al., 2006; Higham et al., 2006b). Generally,
there has been a tendency for bone previously dated using less
rigorous methods to produce substantially different and often older
results when an ultrafiltration pre-treatment is applied (Higham,
2011). We have attributed this to the more successful removal of
contaminants from the collagen. Whilst ultrafiltration will not
remove high molecular weight contamination, on a routine basis it
appears to be a markedly better method than simple Longin
collagen extraction particularly when dating old and poorly
preserved bone.
The first goal of our research at Geißenklösterle was to test
whether the initial series of radiocarbon dates undertaken at the
ORAU from the site were reproducible. This was achieved by
obtaining new material for dating from the same bones and re-
dating them with the application of the additional ultrafiltration
step. Reproducible results would suggest that the original
measurements were reliable, and lend strength to alternative
explanations for the variation in the chronology as outlined above
by Conard and Bolus (2003, 2008) and Zilhão and D’Errico (1999,
2003a, b). On the other hand, significantly different results would
imply a problem with the initial dates, and might require an
explanation based on the differences in the sample pre-treatment
chemistry applied to the paired bone measurements. Only after
clarifying these issues would it be possible to test the models
available for explaining the timing of the advent of innovations
attributed to the Aurignacian.
Several samples were selected for analysis that had yielded AMS
results apparently out of sequence in the chronology. For example,
OxA-6077 and -6076, had resulted in unexpectedly young ages of
32 000 and 33 000 BP (Conard and Bolus, 2003), despite coming
fromAH IIIc, a near-sterile horizon found between the latestMiddle
Palaeolithic and the Lower Aurignacian AH III. Each of the bones
was sampled using an NSK Electer GX drill with a tungsten carbide
drill bit. Ideally, we aimed to take at least 500 mg of bone for
analysis. The bones were pre-treated using the manual Oxford
method (Bronk Ramsey et al., 2004a; Higham et al., 2006b; Brock
et al., 2010), initially with decalcification using 0.5 M HCl,
removal of humates using 0.1 M NaOH, then re-acidification using
0.5 M HCl. Each step was interspersed with ultrapure water rinses.
The samples were gelatinised in water adjusted to pH3 at 75?C for
20 h, and the filtrate recovered using an EzeeFilter?. This gelatin
was further treated by ultrafiltration using a Vivaspin? 30 kDa
Molecular Weight Cut Off (MWCO) ultrafilter (see Bronk Ramsey
et al., 2004a; Higham et al., 2006b). The >30 kDa fraction was
freeze-dried and retained for AMS dating. Samples of ultrafiltered
gelatin are denoted by the prefix ‘AF’ in Table 1.
Samples of pre-treated bone gelatin were combusted in a Carlo-
Erba Elemental Analyser interfaced with a Sercon 20-20 IR mass
spectrometer and a CO2collection unit. The system operates in
continuous flow mode using a He carrier gas. This enables the
measurement of d15N and d13C, nitrogen and carbon concentration,
and C:N atomic ratios on the same fraction of collagen that is dated.
d13C values in this paper are reported with reference to VPDB and
d15N results are reported with reference to AIR. Graphitisation was
byreductionofCO2overanironcatalystinanexcessH2atmosphere
at 560?C (BronkRamseyet al., 2000; Dee and Bronk Ramsey, 2000).
TheOxfordAMSradiocarbon instrumentation isdescribedbyBronk
Ramsey et al. (2000, 2004b). Radiocarbon dates were corrected for
chemistry preparation using a new background subtraction specific
for bone, described in detail in Wood et al. (2010).
Results
The new results are shown in Table 1, alongside all analytical
data we routinely measure. In many instances, they show consid-
erable differences from the previously determined ORAU results
obtained in the 1980s and 1990s using pre-treatment methods that
are no longer applied (ion-exchanged gelatin preparation). The
results reinforce our observations first made in 2001, and since
repeated at numerous sites, that ultrafiltration preparation often
produces older results for Palaeolithic bones compared with other
methods. We have attributed this to an improved contaminant
removal using ultrafiltration (Jacobi et al., 2006; Higham et al.,
2006b), but other factors are also acknowledged as playing a role.
We have an improved background correction and reduced AMS
measurement backgrounds (Bronk Ramsey et al., 2004b; Higham
et al., 2006b). A bone-specific background is incorporated within
our blank correction algorithm (although the primary effect of this
for many samples is to increase the error, rather than the age).
Finally, we also note that for some very low yielding ion-exchanged
gelatin determinations, the possibility of column bleed may have
affected some results by introducing younger laboratory-based
contaminants (Higham et al., 2006b).
In 2008, more bones were sampled from the site using the same
methods described above. These included both cut-marked and
other humanly modified bones that had not been previously
sampled, as well as further samples previously dated at the Oxford
and Kiel laboratories, which appeared out of sequence, or too
young. The details of which types of bone were sampled and the
2In this paper, all ages ‘BP’ or ‘ka BP’ are conventional radiocarbon ages BP.
Calendar/calibrated ages are given with the term ‘cal BP’.
T. Higham et al. / Journal of Human Evolution xxx (2012) 1e13
5
Please cite this article in press as: Higham, T., et al., Τesting models for the beginnings of the Aurignacian and the advent of figurative art and
music: The radiocarbon chronology of Geißenklösterle, Journal of Human Evolution (2012), doi:10.1016/j.jhevol.2012.03.003

Page 6

Table 2
Radiocarbon AMS determinations of ultrafiltered collagen from Geißenklösterle.
OxA Context/LevelSpecies and material datedRadiocarbon
age BP
Error Used
(mg)
Yield
(mg)
%Yld%C
d13C
(&)
C:NComments
Gravettian
21740 GK 99 Ir 185
Ursus spelaeus, parietal
cranium with cutmarks?
Mammuthus primigenius, rib, impact marks
Mammuthus primigenius, rib (?), cut-marked
Rangifer tarandus, metacarpal, impact mark
26,420230394 18.3 4.644.3 ?21.7 3.3
21660 GK 130 It 328
21739 GK 26 Ia 18
21661 GK 86 Ic 122
27,960
28,600
32,900
290 1040
290
450
22.5
43.3
12.2
2.2
7.8
4.1
41.5 ?20.4 3.2
49.6 ?20.6 3.4
44.2 ?18.3 3.1
cf. OxA-5229 (27,950 ? 550)
557
300
cf. OxA-5161 (30,300 ? 750)
Upper Aurignacian
21737 GK 33 IIa 80
21656 GK IIa 131
21738 GK IIb 143
cf. Rangifer tarandus
Equus ferus, scapula
Equus ferus, humerus
(retouch marks inferred)
Equus ferus, humerus
Ursus spelaeus, rib fragment
with a cutmark
Mammuthus primigenius,
rib fragment with impact point
35,700
33,000
34,900
650
500
600
630
520
602.6 20.4
30.3
27.4
4.8
5.3
3.4
46.0 ?21.0 3.3
43.4 ?20.2 3.2
46.0 ?20.7 3.3
cf. OxA-5707 (33,200 ? 800)
21742 GK 67 IIb 931
21727 GK 57 IIb 706
34,800
34,100
600
550
528
440
23.5
37.9
4.5
8.6
44.9 ?20.3 3.3
47.2 ?20.7 3.3
cf. KIA-8958 (31,870 þ 260/?250)
21724 GK 58 IIb 24633,950550 64016.42.645.4 ?21.0 3.3
cf. KIA 8960 (29,800 ? 240)
Early or lower Aurignacian
21725 GK 86 III 294Large unidentified mammal rib fragment
(cf. Coelodonta antiquitatus/
Mammuthus primigenius)
Equus ferus, humerus (retouched)
Rangifer tarandus, tibia, impact mark
Coelodonta antiquitatus,
humerus, no cutmarks
37,40080056024.44.446.4 ?20.4 3.3
cf. KIA 8963 (31,180 þ 270/?260)
21726 GK 55 IId 319
21659 GK 77 III 627
21744 GK 77 III 641
34,200
35,050
36,850
550
600
750 1071
620
480
13.2
12.1
17.0
2.1
2.5
1.6
46.3 ?19.5 3.3
44.1 ?18.9 3.2
45.4 ?19.4 3.3
cf. OxA-6256 (30,100 ? 550)
cf. OxA-6255 (32,900 ? 850)
21745 GK 66 IIIa 1073 Rangifer tarandus, tibia, impact marks
21746 GK 67 IIIa 1453 Rangifer tarandus, tibia with cutmarks
21722 GK 66 III 1144
Equus ferus, distal femur
21743 GK 67 IIIb 1655 Rangifer tarandus, tibia with
impact and cutmarks
21723 GK 69 IIIb 958Artiodactyl limb bone fragment
21721 GK 57 IIIb 1238 cf. Coelodonta antiquitatus/Mammuthus
primigenius bone fragment with scrape
marks, i.e, humanly-modified bone.
36,650
36,850
38,900
36,100
750
800
1000
700 1060
770
690
530
42.3
51.8
27.8
72.0
5.5
7.5
5.2
6.8
45.2 ?19.0 3.4
45.4 ?18.7 3.4
46.8 ?19.6 3.3
46.6 ?18.6 3.4
cf. KIA13074 (34,800 þ 290/?280)
cf. KIA13075 (34,330 þ 310/?300)
cf. OxA-4595 (40,200 ? 1600)
cf. KIA13076 (34,080 þ 300/?290)
37,800
37,300
900
800
580
580
28.1
8.1
4.8
1.4
47.4 ?19.1 3.4
43.6 ?19.6 3.3
cf. KIA8959 (34,220 þ 310/?300)
cf. KIA 8962 (28,640 þ 380/?360)
Sterile level
21657 GK 57 IIIc 2430 Cervus elaphus, tibia, no human modification
21658 GK 57 IIIc 2389 Capra ibex, left tibia, no human modification
39,400
38,300
1100
900
480
420
20.4
12.6
4.2
3.0
43.9 ?19.4 3.1
44.2 ?18.3 3.1
cf. OxA-6076 (33,600 ? 1900)
cf. OxA-6077 (32,050 ? 600)
Middle Palaeolithic
21720 GK 78 IV 1495cf. Ursus spelaeus, juvenile shaft
fragment. Possible impact.
Capra ibex, phalanx I, which articulates with
metataurus. No clear cutmarks
although two are inferred.
35,500650 64014.0 2.2 46.9 ?20.7 3.3
cf. KIA19556 (37,780 þ 520/?490)
21741 GK 48 VII 456
48,600 3200478 38.86 8.148.3 ?18.7 3.3
See Table 1 caption for details of the analytical data in the table. In the comments column previous dates on the same bone specimen are shown. See text for details.
Table 1
Initial radiocarbon determinations from Geißenklösterle, Germany designed to test the reliability of previous outlying results that were inconsistent with the stratigraphic
sequence.
OxA-numberSample identification Species Radiocarbon age BPUsed (mg) Yield (mg)%Yld. Coll. %C
d13C (&)
?20.5
?20.2
?22.3
?19.4
?19.4
?18.3
?19.1
?18.9
?20.4
?20.4
?19.1
?18.3
C:N
5707 (AI)
21656 (AF)
6076 (AI)
21657 (AF)
6077 (AI)
21658 (AF)
6256 (AI)
21659 (AF)
5229 (AI)
21660 (AF)
5161 (AI)
21661 (AF)
GK IIa 131
Equus ferus, scapula 33,200 ? 800
33,000 ? 500
33,600 ? 1900
39,400 ? 1100
32,050 ? 600
38,300 ? 900
30,100 ? 550
35,050 ? 600
27,950 ? 550
27,960 ? 290
30,300 ? 750
32,900 ? 450
720
520
500
480
520
420
560
480
1000
1040
500
300
29.7
27.4
20.3
20.4
12.6
12.6
4.9
12.1
11.3
22.5
8.7
12.2
4.1
5.3
4.1
4.2
2.4
3.0
0.9
2.5
1.1
2.2
1.7
4.1
43.8
43.4
40.8
43.9
39.6
44.2
41.3
44.1
42.0
41.5
41.7
44.2
3.2
GK 57 IIIc 2430
Cervus elaphus, tibia3.1
GK 57 IIIc 2389
Capra ibex, left tibia3.1
GK 77 III 627
Rangifer tarandus, tibia 3.2
GK 130 It 328
Mammuthus primigenius, rib 3.2
GK 86 Ic 122
Rangifer tarandus, metacarpal3.1
There are two determinations from each bone, an AI (ion-exchanged gelatin treatment) and AF (gelatinisation and ultrafiltration) (note that previously these initial dates were
identified incorrectly as being AG or gelatin treated determinations in Higham, 2011). Stable isotope ratios are expressed in & relative to vPDB with a mass spectrometric
precision of ?0.2&. Yield represents the weight of collagen or ultrafiltered collagen in milligrams. %Yld is the percent yield of extracted collagen as a function of the starting
weight of the bone analysed. %C is the carbon present in the combusted collagen. C:N is the atomic ratio of carbon to nitrogen and is acceptable if it ranges between 2.9 and 3.5.
T. Higham et al. / Journal of Human Evolution xxx (2012) 1e13
6
Please cite this article in press as: Higham, T., et al., Τesting models for the beginnings of the Aurignacian and the advent of figurative art and
music: The radiocarbon chronology of Geißenklösterle, Journal of Human Evolution (2012), doi:10.1016/j.jhevol.2012.03.003

Page 7

results are shown in Table 2. Again, some significant differences
between thepreviously determined
measurements are noted from both laboratories. A large mammal
rib dated in the Kiel laboratory, for example (KIA-8962), had
previously produced an age of 28,640 þ 380/?360 BP. A direct
redate on the same specimen produced an age of 37,400 ? 800 BP.
An age of w28,600 BP for the Early Aurignacian was always
considered a surprisingly young result but determining whether it
was due to site taphonomy, sample contamination or radiocarbon
variability was not possible until now. The research presented here
suggests strongly that the problems are to do with the original AMS
measurements, and may be related to the pre-treatment chemistry
applied, with the most likely reasons being ultrafiltration or the
lack of application of a base wash to some of the samples. The
background correction mentioned above is probably not the prin-
cipal reason for divergent results, because the size of the correction
would not be sufficiently large enough to account for the differ-
ences between the two.
resultsand thenew
The difference between measurements is the key feature of the
series of AMS dates, both old and new. Fig. 4 summarises all of the
results, with ultrafiltered determinations in red and the original
non-ultrafiltered dates in black. Ultrafiltered conventional radio-
carbon ages are statistically significantly older than their paired
non-ultrafiltered dates in 14 out of 17 cases when tested using the
error-weighted mean method of Ward and Wilson (1978). Reduced
offsets between dates obtained with and without ultrafiltration are
apparent in the upper levels, in the Gravettian, and this is what we
would expect to see if contamination effects are the reason for the
differences between ultrafiltered and non-ultrafiltered samples.
Calibration and interpretation
The INTCAL09 interim calibration dataset (Reimer et al., 2009) is
the result of the integration of several different datasets, including
those from the Cariaco basin foraminifera (Hughen et al., 2004,
2006) and Barbadian corals (Fairbanks et al., 2005), augmented
by other forams from a core taken along the Iberian Margin (Bard
et al., 2004). We used this INTCAL09 curve to produce a revised
chronology for Geißenklösterle, based on the new ultrafiltered
determinations described above. It is important to note that
through this section of the INTCAL09 record, the dataset is based on
marine records and there are some uncertainties regarding it and
the question of constant reservoir offsets to the atmosphere.
Further work is underway to test this, and for the purpose of this
discussion the calibration of radiocarbon between 26,000 and
50,000 BP should be considered a preliminary interpretation. The
model produced is shown in Fig. 5. We used a Bayesian modelled
sequence generated with OxCal 4.1 (Bronk Ramsey, 2001, 2009a).
We compare the modelled results against a climate proxy, in this
case the Hulu-tuned (after Weninger and Jöris, 2008) GICC05 d18O
ice core data published by Svensson et al. (2006) and Andersen
et al. (2006). Further work is required to determine the synchro-
nicity of the climatic changes recorded in the Greenland ice and
those of northern Europe (Blaauw et al., 2010). While this part of
our analysis may require some modification in the future, it is
expected that the overall pattern of the chronology will remain
very similar, and is therefore worth presenting now.
The model priors consist of the stratigraphic sequence informa-
tion obtained from the excavation. A uniform distribution of all
radiocarbon ages is assumed within each successive archaeological
phase in the model. Between each phase boundary distributions are
modelledthatcorrespondwiththebeginningandendofeachperiod
ofactivity.Thoughundatedbyradiocarbon,theseprovideimportant
probability distribution functions (PDFs) representing the start and
end dates of these phases and, as such, are particularly useful. The
model is a conservative one, following the most recently published
chrono-stratigraphic frameworks (Conard and Bolus, 2003, 2008).
The outlier detection method of Bronk Ramsey (2009b) was
used to analyse the radiocarbon data within the constraints of the
sequence. A t-type general outlier model with prior probability set
at 0.05 was employed. This enabled the probability of the individ-
ual’s likelihood being outliers within the model to be assessed. The
results of the first iteration of the modelling are shown in Table 3,
the left side of the table shows the unmodelled calibrations and the
italicised data on the right hand side shows the modelled ranges at
68.2% and 95.4% probability ranges. All results are rounded to the
nearest decade. The results of the outlier analysis are shown in
Table 4. There is one demonstrable outlier within the sequence of
24 AMS dates (OxA-21720) (statistically speaking, one outlier in 20
should be expected). This result appears to be too young for its
stratigraphic position. The model was run several times to assess
reproducibility and found to be acceptable. Convergence values for
the likelihoods in the model ranged between 99.4 and 100%. This
Figure 4. Comparison data for ultrafiltered (red) and non-ultrafiltered (black) deter-
minations from Geißenklösterle. (For interpretation of the references to colour in this
figure legend, the reader is referred to the web version of this article.)
T. Higham et al. / Journal of Human Evolution xxx (2012) 1e13
7
Please cite this article in press as: Higham, T., et al., Τesting models for the beginnings of the Aurignacian and the advent of figurative art and
music: The radiocarbon chronology of Geißenklösterle, Journal of Human Evolution (2012), doi:10.1016/j.jhevol.2012.03.003

Page 8

provides some testof theeffectivenessof the MCMC algorithm used
in the modelling and the degree to which a representative solution
has been found for a specific parameter or likelihood in the model.
Ideally, values ought to be >95%.
The radiocarbon determinations obtained include bone samples
previously dated, but which had no clear human modification.
Compared with other Middle to Upper Palaeolithic sites sampled
as part of our wider project,3
the bone assemblage from
Geißenklösterle had few bones with clear cutmarks that were
easily identifiable without a microscope. This lack of visibility may
be related to the rather poor surface preservation state of the bone.
Nonetheless, with the exception of cave bear, the vast majority of
the faunal assemblage is thought to be anthropogenically-derived
on the basis of species distribution patterns, the direct association
with lithic assemblages and other indicators of hominin activity in
the bone assemblage. All new samples selected for dating had
convincing evidence of human modification to ensure that the
resulting dates documented the presence of humans at the site.
It is important to consider the effects of these data on the age
modelling, so we tested the effect of the non-modified bone
Figure 5. Bayesian model for the Geißenklösterle sequence produced using OxCal 4.1 (Bronk Ramsey, 2001). The radiocarbon ages are compared against the interim INTCAL09
dataset of Reimer et al. (2009). The model is based on the series of separate excavated phases which are divided by boundaries. Individual radiocarbon likelihoods are shown by the
light shaded distributions, whilst the darker outlines represent posterior probability distributions. Figures in brackets next to the OxA-numbers represent outlier detection results
(first value is a posteriori and second value is a priori outlier value). The calibrated ages are compared tentatively with the NGRIP d18O record, with Greenland interstadials
numbered where relevant (data from Andersen et al., 2006; Svensson et al., 2006, 2008). It is important to note that the comparison of the age model for INTCAL09 and the NGRIP
age model is tentative and there is uncertainty over whether the climatic changes recorded in Greenland are completely synchronous with mainland Europe (see Blaauw et al.,
2010).
3This paper forms part of the NERC funded ‘Dating of the Middle-Upper Palae-
olithic transition in western Europe using ultrafiltration AMS radiocarbon’ project.
T. Higham et al. / Journal of Human Evolution xxx (2012) 1e13
8
Please cite this article in press as: Higham, T., et al., Τesting models for the beginnings of the Aurignacian and the advent of figurative art and
music: The radiocarbon chronology of Geißenklösterle, Journal of Human Evolution (2012), doi:10.1016/j.jhevol.2012.03.003

Page 9

determinations on the results by increasing the prior outlier
probabilities for non-modified bones in the sequence and running
the model again. We increased the outlier priors from 0.05 up to
a final value of 1.00 (indicating that they were 100% likely to be
outliers). The results suggested that there is a small effect on the
boundary distribution for the AH III start boundary (the boundary
representing the start of the Early Aurignacian). Fig. 6 shows the
results of modelling the AH III start boundary when all of the
determinations are included, and when only humanly-modified or
cut bone material is used. When all bones are included, the
boundary distribution for the start of AH III is 42,940e42,180 BP
(68.2% prob.) and 43,410e41,860 BP (95.4%). When only cut bone
is considered, the boundary is 42,640e41,900 BP (68.2%) and
43,060e41,480 BP (95.4%). The modelled results suggest, then, that
when humanly-modified material is specified in the model, the age
ranges are slightly younger than they otherwise would be. We may
quantify this by calculating the difference between the two prob-
ability distributions for the start of AH III. This shows that the age
effect is 250 to ?800 years (68.2% prob.). Previous zooarchaeo-
logical studies have shown that the vast majority of mammalian
bone from the lower Aurignacian is of anthropogenic origin
(Münzel and Conard, 2004). The only significant exception is the
presence of cave bear bones, and this vegetarian species is not
a bone collector.
Previous radiocarbon determinations from the lower part of
the Middle Palaeolithic sections of the site (AH VII and AH VIII)
had resulted in a maximum age of w41,400 BP (KIA-19560). We
tested whether this was also an underestimate of the lower parts
of the sequence. A sample of bone from AH VII was selected for
dating. The sample was a first phalanx of a Capra ibex, which was
found in articulation. Selecting articulated bones reduces the
possibility that the material is reworked or intrusive because it is
likely to have been in anatomical position when deposited. We
obtained a result of 48,600 ? 3200 BP (OxA-21741), suggesting
that the bone previously dated from this level of 31,620 þ 391/
?373 BP (KIA-19557) and the three non-ultrafiltered dates of
bone from the level below this (KIA-19558: 33,430 þ 480/
?450 BP, KIA-19561: 40,090 þ 640/?600 BP and KIA-19560:
41,410 þ 1500/?1260 BP) are significant underestimates, if we
can rely on their context securely within this level. The result
suggests that bone from below AH VII is close to the radiocarbon
limit of w50,000 BP recently re-estimated for ORAU (Wood et al.,
2010), but not beyond it. The possibility that AH VII is >50,000 BP,
however, would require testing with additional samples to
increase confidence in the single determination that we have
obtained.
Richter et al. (2000) have presented results of TL measurements
on burnt flint from AH III and ESR measurements on teeth from AH
IV. We compared these values with our determinations in
a Bayesian model. The results show good agreement (Fig. 7). This
model also includes 100% outlier probabilities for all non-cut or
modified bone as described previously.
Table 3
Calibrated data for the Geißenklösterle radiocarbon ages and model as determined using OxCal.
Calibrated age range (68.2% prob.)Calibrated age range (95.4% prob.) Modelled range (68.2% prob.)Modelled range (95.4% prob.)
From ToFrom To From ToFrom To
End boundary
OxA-21740
OxA-21660
OxA-21739
OxA-21661
31,150
31,220
32,570
33,500
37,760
29,690
30,930
31,650
32,470
36,710
31,350
31,340
33,050
34,430
38,580
26,610
30,650
31,450
31,920
36,470
31,210
32,560
33,490
38,370
30,900
31,660
32,490
36,870
31,310
33,000
34,410
38,680
30,620
31,470
31,940
36,620
Gravettian
End Aurignacian
OxA-21656
OxA-21738**
OxA-21737
OxA-21724**
OxA-21727**
OxA-21742
38,780
39,080
40,150
40,490
39,760
39,830
40,100
37,540
37,940
39,040
39,310
38,530
38,640
39,000
39,530
40,050
40,740
41,030
40,370
40,450
40,700
36,910
37,460
38,710
38,870
37,930
38,070
38,660
38,410
40,700
41,560
39,710
40,050
40,560
37,060
39,220
40,220
38,000
38,510
39,090
38,820
41,270
41,990
40,500
40,650
41,230
36,610
38,740
39,280
37,290
37,500
38,670
AH II
Transition AH III/AH II
OxA-21726**
OxA-21744
OxA-21725
OxA-21746**
OxA-21745
OxA-21722
OxA-21743**
OxA-21721**
OxA-21659**
OxA-21723
41,000
41,270
42,190
42,400
42,210
42,120
42,660
41,890
42,360
41,430
42,500
39,910
40,110
41,300
41,550
41,280
41,190
41,860
40,910
41,500
40,420
41,650
41,540
42,230
42,640
42,810
42,660
42,570
43,150
42,300
42,790
41,950
42,920
39,370
39,670
40,790
41,080
40,740
40,650
41,370
40,350
41,030
39,890
41,170
40,060
42,340
42,740
42,380
42,220
44,040
41,850
42,680
40,910
43,090
38,610
41,230
41,600
41,200
41,090
42,510
40,610
41,530
39,430
41,780
40,760
42,990
43,480
43,080
42,860
44,860
42,370
43,420
41,370
44,040
37,670
40,590
41,040
40,490
40,370
41,910
39,610
40,970
38,820
41,270
AH III
Transition Sterile/AH IIIc
OxA-21658
OxA-21657
42,940
43,300
43,410
42,180
42,390
42,400
43,410
43,890
44,100
41,860
42,040
42,040
43,450
44,410
42,080
42,770
44,330
45,390
41,630
42,110
Sterile
Transition AH IV/Sterile
OxA-21720
43,830
44,460
42,450
42,510
45,040
46,380
42,050
42,100
41,43039,97041,80039,070
IV
Transition AH V/AH IV
45,48042,59048,470
In the right hand columns (in italics), the posterior distributions are shown, at 68.2% and 95.4% probability. The data in this table is illustrated in Fig. 5. ** These OxA-numbers
denote cut-marked or humanly-modified samples from the Aurignacian levels.
T. Higham et al. / Journal of Human Evolution xxx (2012) 1e13
9
Please cite this article in press as: Higham, T., et al., Τesting models for the beginnings of the Aurignacian and the advent of figurative art and
music: The radiocarbon chronology of Geißenklösterle, Journal of Human Evolution (2012), doi:10.1016/j.jhevol.2012.03.003

Page 10

Synthesis
How do the results for the Early Aurignacian at Geißenklösterle
fit with other dated examples where we find a similar industry? At
present, it is difficult to compare with confidence because most of
the other key sites have been dated using less refined pre-
treatment techniques, and we think many of the results need to
be tested. There is, however, recent ultrafiltered collagen and
ABOx-SCeprepared charcoal results from the Italian site of Grotta
di Fumane (Higham et al., 2009; Higham, 2011), where we find an
important Proto-Aurignacian level (A2). Levels of the same industry
have also been dated recently at Riparo Mochi (see Douka et al.,
2012), the French site of Arcy-sur-Cure (Higham et al., 2010) and
the Catalan site of Abric Romaní (Camps and Higham, 2012) (Fig. 6).
We mayalso include the Early Aurignacian levels at the Abri Pataud
(Higham et al., 2011) and La Cottes (Talamo et al., 2012), where
extensive ultrafiltered series are available. We leave out the recent
dates from Isturitz, however, because they were not ultrafiltered
and they have extremely large standard errors (Szmidt et al., 2010).
We may compare the start boundaries denoting the beginning
of these respective levels at the various sites (Fig. 6). These
boundaries are based on a large number of modelled likelihoods
and ought to be robust. The results show some variation in the start
date ranges and suggest overall that there is a considerable time-
span in the age of the Early Aurignacian across Europe. What is
apparent, however, is that the Geißenklösterle results are demon-
strably earlier than all of the other sites (Fig. 6).
Further chronometric comparisons are possible for Aurignacian
sites in areas where deposits from the Campanian Ignimbrite (CI)
tephra are located. The CI was probably erupted from the Phlegrean
Fields (near Naples, Italy) (Giaccio et al., 2006, 2008) and the tephra
from it covers most of the Italian Peninsula as well parts of western
Russia, Greece and the eastern Mediterranean (Fedele et al., 2008).
Sites dating to the Aurignacian and Ulluzzian sealed beneath the
tephra therefore pre-date it. The age of the ash has been deter-
mined using40Ar/39Ar techniques at 39,280 ? 110 yr BP (De Vivo
et al., 2001; Pyle et al., 2006). A glance at Fig. 6 shows that all of
the Aurignacian sites in our comparison sit prior to the date of the
CI tephra and therefore are consistent with the chronological
picture from these other sites. In addition, the results from Gei-
ßenklösterle demonstrate that, in terms of climate and environ-
ment, the Aurignacian occupations pre-date the Heinrich 4 event
by several millennia, if our comparisons against NGRIP are valid.
The majority of scholars conclude that the Aurignacian is the
earliest signature of the first modern humans in Europe. Recent
research suggests that this is not likely to be the case. Benazzi et al.
(2011) have shown that the Uluzzian of Italy and Greece is likely to
be a modern human industry based on the reanalysis of infant teeth
in the archaeological site of Cavallo, and also demonstrated that it
dates to 45,000e43,000 cal BP. Other dated examples from other
Uluzzian sites (e.g., Higham et al., 2009) fall into the same period,
and the Uluzzian is always stratigraphically below the Proto-
Figure 6. PDFs for the start boundary of AH III derived from two Bayesian models, the first of which includes all of the bone determinations and the second which includes only cut
and humanly-modified bones from the Geißenklösterle. These PDFs are compared with other boundary distributions modelled for the start of the Early Aurignacian or Proto-
Aurignacian at several other sites in Europe (Fig. 1 for locations).
Table 4
Prior and posterior outlier data.
ElementPrior Posterior
OxA-21720
OxA-21657
OxA-21658
OxA-21723
OxA-21659
OxA-21721
OxA-21743
OxA-21722
OxA-21745
OxA-21746
OxA-21725
OxA-21744
OxA-21726
OxA-21742
OxA-21727
OxA-21724
OxA-21737
OxA-21738
OxA-21656
OxA-21661
OxA-21739
OxA-21660
OxA-21740
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
87
5
4
5
6
4
4
8
4
4
4
4
14
4
4
4
5
4
7
5
5
4
4
There is one outlier of significance amongst the new AMS series published in this
paper (OxA-21720).
T. Higham et al. / Journal of Human Evolution xxx (2012) 1e13
10
Please cite this article in press as: Higham, T., et al., Τesting models for the beginnings of the Aurignacian and the advent of figurative art and
music: The radiocarbon chronology of Geißenklösterle, Journal of Human Evolution (2012), doi:10.1016/j.jhevol.2012.03.003

Page 11

Aurignacian in Italian sites where both co-occur. This adds an
additional level of complexity to the emerging picture of early
human dispersals and suggests that the Aurignacian does not
represent the earliest evidence of our species in Europe. Further
recent research has demonstrated that anatomically modern
humans also attained the far reaches of northwestern Europe at
a similarly early period. Higham et al. (2011) have shown that KC4,
the anatomically modern human maxilla from the Kent’s Cavern
site in England, dates to w43,000e41,000 cal BP. Although not
associated with an archaeological industry, the age of this specimen
shows it to be earlier than any other directly-dated modern human
bone in northwestern Europe. The start of the Aurignacian at Gei-
ßenklösterle, as described above, dates to 43,060e41,480 cal BP (at
95.4% prob.), so is directly comparable with the KC4 modelled age.
Taken together, these results suggest that modern humans
arrived in Europe as early as w45,000 cal BP and spread rapidly
across Europe to as far as southern England between 43,000 and
41,000calBP.ThedatesforthelowerAurignacianatGeißenklösterle
fallinthesame period andappearto pre-date theages fortheProto-
Aurignacian and Early Aurignacian in other regions (Fig. 6). The new
results suggest that the caves of the Swabian Jura document the
earliest phase of the Aurignacian, and the region can be viewed as
one of the key areas in which a variety of cultural innovations,
including figurative art, mythical images, and musical instruments,
are first documented. These dates are consistent with the Danube
Valley serving as an important corridor for the movement of people
and ideas (Conard, 2002; Conard and Bolus, 2003). The early radio-
carbon dates for the Swabian Aurignacian suggest that this region
contributedtotheevolutionofsymbolicartifactsassociatedwiththis
period, as hypothesised by the Kulturpumpe model. This model is
readily refutable if examples of figurative art, mythical imagery,
three-dimensionally formed personal ornaments and musical
instruments are documented at an earlier date in other regions.
While we find it highly unlikely that all of these innovations have
a unique monocentric origin in the Upper Danube region (Conard,
2008), the available dates are consistent with this variant of what
canbecalledthe‘strong’Kulturpumpemodel.Weare,however,more
sympathetictowhatcouldbecalledthe‘weak’kulturpumpemodelin
Figure 7. Modelled resultsforAH IIand AH IIIshowingcomparison against the TL and ESRmean agesobtained previouslybyRichteret al.(2000).The TL/ESR agesaregiven inred,the
darkerredistheposteriorprobabilityormodelleddistribution,whilstthelighterredistheoriginallikelihood(unmodelled).Thenon-humanlymodifiedbonedeterminationsaregiven
a 100% likely outlier probability in this model, as described in the text, so this model is only based on material which we are confident was deposited by humans. See Fig. 5 caption for
details of other parts of the model and climate comparison.
T. Higham et al. / Journal of Human Evolution xxx (2012) 1e13
11
Please cite this article in press as: Higham, T., et al., Τesting models for the beginnings of the Aurignacian and the advent of figurative art and
music: The radiocarbon chronology of Geißenklösterle, Journal of Human Evolution (2012), doi:10.1016/j.jhevol.2012.03.003

Page 12

which,paralleltotheinnovationsdocumentedintheSwabianCaves,
other regions also contributed analogous cultural innovations
indicative of a stronger reliance on symbolic artifacts than in the
Middle Paleolithic and Early Upper Paleolithic.
Conclusions
A new series of AMS determinations reported here show that
the previous chronology of Geißenklösterle was dominated by
erroneous radiocarbon measurements. These dates fuelled an
extensive debate on the chronology of the earliest European
Aurignacian and the integrity of the archaeological sequence at the
site, which remained unresolved for many years. The new deter-
minations obtained using ultrafiltration pre-treatment on samples
of bone appear coherent and reproducible. They show agreement
with previous non-AMS ages derived using TL dating of burnt flints
and ESR dates on fossil teeth. For the crucial lower Aurignacian of
AH III, the new dates suggest that many of the previous radio-
carbon results published by Conard and Bolus (2003, 2008) are
aberrant, and mostly too young. Some, however, by virtue of their
poor precision, remain in agreement with the newly modelled
results. When we compare old and new dates, our analysis shows
that 14 of 17 determinations are statistically distinguishable, and
that the ultrafiltered collagen fraction is older in all cases but one.
In the lower Mousterian sections of the site, all previous dates
appear to underestimate the age by about 8e20,000 radiocarbon
years and this is mirrored in the revised ages for some of the other
previous dates from the Aurignacian levels. In more modern
material, the effect of contaminants is reduced, but in the deeper
sections, where14C activities are <2e3% of modern, their effect
can be much more dramatic as our new results show. This mirrors
results from other sites in Palaeolithic Europe that we have rean-
alysed using ultrafiltration and ABOx-SC methods (Jacobi et al.,
2006; Higham et al., 2006b; Douka et al., 2010; Higham, 2011)
and emphasises once more that building and refining chronologies
for this period must remain a major research goal for the imme-
diate future. It will only become possible to compare different sites
across time and space with confidence when we have a trusted
chronology.
The new radiocarbon dates from Geißenklösterle document the
presence of the Aurignacian in the Swabian Jura prior to the
Heinrich 4 cold phase, with the EarlyAurignacian beginning around
42,500 cal BP. In the coming years, excavations in the Swabian Jura
will continue and new radiometric dates should contribute to an
improved understanding of the spatial-temporal development of
the Aurignacian and its innovative material culture.
Acknowledgements
This research was funded by a NERC Standard grant (NE/
D014077/1) as part of the project ‘Dating of the Middle-Upper
Palaeolithic transition in western Europe using ultrafiltration AMS
radiocarbon’ for which we are extremely grateful. R. Wood was
funded by a tied studentship to this grant. We are very grateful to
M. Bolus, S. Münzel and M. Malina (Tübingen) for their assistance
and to the staff of the Oxford Radiocarbon Accelerator Unit (ORAU)
for their careful laboratory work. The manuscript benefitted from
the work of three referees.
Appendix A. Supplementary data
Supplementary data related to this article can be found online at
doi:10.1016/j.jhevol.2012.03.003.
References
Andersen,K.K.,Svensson,A.,Johnsen,S.J.,Rasmussen,S.O.,Bigler,M.,Rothlisberger,R.,
Ruth, U., Siggaard-Andersen, M.-L., Steffensen, J.P., Dahl-Jensen, D., Vinther, B.M.,
Clausen, H.B., 2006. The Greenland ice core chronology 2005, 15e42 ka. Part 1:
constructing the time scale. Quatern.Sci. Rev. 25, 3246e3257.
Bard, E., Ménot-Combes, G., Rostek, F., 2004. Present status of radiocarbon cali-
bration and comparison records based on Polynesian corals and Iberian Margin
sediments. Radiocarbon 46, 1189e1202.
Beck, J.W., Richards, D.A., Edwards, R.L., Silverman, B.W., Smart, P.L., Donahue, D.J.,
Hererra-Osterheld, S., Burr, G.S., Calsoyas, L., Jull, A.J.T., Biddulph, D., 2001.
Extremely large variations of atmospheric14C concentration during the last
glacial period. Science 292, 2453e2458.
Benazzi, S., Douka, K., Fornai, C., Bauer, C.C., Kullmer, O., Svoboda, J., Pap, I.,
Mallegni, F., Bayle, P., Coquerelle, M., Condemi, S., Ronchitelli, A., Harvati, K.,
Weber, G.W., 2011. Early dispersal of modern humans in Europe and implica-
tions for Neanderthal behaviour. Nature 479, 525e528.
Blaauw, M., Wohlfarth, B., Andrés Christen, J., Ampel, L., Veres, D., Hughen, K.A.,
Preusser, F., Svensson, A., 2010. Were last glacial climate events simultaneous
between Greenland and France? A quantitative comparison using non-tuned
chronologies. J. Quatern. Sci. 25, 387e394.
Bolus, M., Conard, N.J., 2001. The late middle paleolithic and earliest upper Paleo-
lithic in central Europe and their relevance for the out of Africa hypothesis.
Quatern. Int. 75, 29e40.
Brock, F., Higham, T.F.G., 2009. AMS radiocarbon dating of Paleolithic-aged charcoal
from Europe and the Mediterranean rim using ABOX-SC. Radiocarbon 51,
839e846.
Brock, F., Higham, T.F.G., Ditchfield, P., Bronk Ramsey, C., 2010. Current pretreatment
methods for AMS radiocarbon dating at the Oxford Radiocarbon Accelerator
Unit (ORAU). Radiocarbon 52, 103e112.
Bronk Ramsey, C., 2001. Development of the radiocarbon calibration program
OxCal. Radiocarbon 43, 355e363.
Bronk Ramsey, C., 2009a. Bayesian analysis of radiocarbon dates. Radiocarbon 51,
337e360.
Bronk Ramsey, C., 2009b. Dealing with outliers and offsets in radiocarbon dating.
Radiocarbon 51, 1023e1045.
Bronk Ramsey, C., Pettitt, P.B., Hedges, R.E.M., Hodgins, G.W.L., Owen, D.C., 2000.
Radiocarbon dates from the Oxford AMS system: archaeometry datelist 30.
Archaeometry 42, 459e479.
Bronk Ramsey, C., Higham, T.F.G., Bowles, A., Hedges, R., 2004a. Improvements to
the pre-treatment of bone at Oxford. Radiocarbon 46, 155e163.
Bronk Ramsey, C., Higham, T., Leach, P., 2004b. Towards high-precision AMS:
progress and limitations. Radiocarbon 46, 17e24.
Brown, T.A., Nelson, D.E., Vogel, J.S., Southon, J.R., 1988. Improved collagen extrac-
tion by modified Longin method. Radiocarbon 30, 171e177.
Camps, M., Higham, T.F.G., 2012. Chronology of the Middle to upper Palaeolithic
transition at Abric Romaní, Catalunya. J. Hum. Evol. 62, 89e103.
Conard, N.J., 2002. The timing of cultural innovations and the dispersal of modern
humans in Europe. Terra Nostra 2002, 82e94.
Conard, N.J. (Ed.), 2006. When Neanderthals and Modern Humans Met. Kerns
Verlag, Tübingen.
Conard, N.J., 2008. A critical view of the evidence for a southern African origin of
behavioral modernity. S. Afr. Archaeol. Soc. Goodwin Ser. 10, 175e179.
Conard, N.J., 2009. A female figurine from the basal Aurignacian of Hohle Fels cave
in southwestern Germany. Nature 459, 248e252.
Conard, N.J., Bolus, M., 2003. Radiocarbon dating the appearance of modern
humans and timing of cultural innovations in Europe: new results and new
challenges. J. Hum. Evol. 44, 331e371.
Conard, N.J., Bolus, M., 2006. The Swabian Aurignacian and its place in European
prehistory. In: Bar-Yosef, O., Zilhao, J. (Eds.), 2006. Toward a Definition of the
Aurignacian. Trabalhos de Arcqueologia, vol. 45. Instituto Português de
Arqueologia, Lisbon, pp. 209e237.
Conard, N.J., Bolus, M., 2008. Radiocarbon dating the late Middle Paleolithic and the
Aurignacian of the Swabian Jura. J. Hum. Evol. 55, 886e897.
Conard, N.J., Malina, M., 2003. Abschließende Ausgrabungen im Geißenklösterle bei
Blaubeuren, Alb-Donau-Kreis. Archäologische Ausgrabungen in Baden-Würt-
temberg 2002, 17e21.
Conard, N.J., Dippon, G., Goldberg, P., 2003. Chronostratigraphy and archaeological
context of the Aurignacian deposits at Geißenklösterle. In: Zilhao, J., d’Errico, F.
(Eds.), The Chronology of the Aurignacian and the Transition Technocomplexes.
Dating, Stratigraphies and Cultural Implications. Trabalhos de Arqueologia, vol.
33. Instituto Português de Arqueologia, Lisbon, pp. 165e176.
Conard, N.J., Malina, M., Münzel, S.C., 2009. New flutes document the earliest
musical tradition in southwestern Germany. Nature 460, 737e740.
D’Errico, F., 2003. The invisible frontier. A multiple species model for the origin of
behavioral modernity. Evol. Anthropol. 12, 188e202.
D’Errico, F., Zilhão, J., Julien, M., Baffler, D., Pelegrin, J., 1998. Neanderthal accul-
turation in western Europe? A critical review of the evidence and its inter-
pretation. Curr. Anthropol. 39, S1eS44.
De Vivo, B., Rolandi, G., Gans, P.B., Calvert, A., Bohrson, W.A., Spera, F.J., Belkin, H.E.,
2001. New constraints on the pyroclastic eruptive history of the Campanian
volcanic Plain (Italy). Miner. Petrol. 73, 47e65.
Dee, M., Bronk Ramsey, C., 2000. Refinement of graphite target production at ORAU.
Nucl. Instrum. Meth. B 172, 449e453.
T. Higham et al. / Journal of Human Evolution xxx (2012) 1e13
12
Please cite this article in press as: Higham, T., et al., Τesting models for the beginnings of the Aurignacian and the advent of figurative art and
music: The radiocarbon chronology of Geißenklösterle, Journal of Human Evolution (2012), doi:10.1016/j.jhevol.2012.03.003

Page 13

Douka, K., Higham, T., Sinitsyn, A., 2010. The influence of pretreatment chemistry on
the radiocarbon dating of Campanian Ignimbrite-aged charcoal from Kostenki
14 (Russia). Quatern. Res. 73, 583e587.
Douka, K., Grimaldi, S., Boschian, G., del Lucchese, A., Higham, T.F.G., 2012. A new
chronostratigraphic framework for the Upper Palaeolithic of Riparo Mochi
(Italy). J. Hum. Evol. 62, 286e299.
Fairbanks, R.G., Mortlock, R.A., Chiu, T.-C., Cao, L., Kaplan, A., Guilderson, T.P.,
Fairbanks, T.W., Bloom, A.L., Grootes, P.M., Nadeau, M.-J., 2005. Radiocarbon
calibration curve spanning 0e50,000 years BP based on paired230Th/234U/238U
and14C dates on pristine corals. Quatern. Sci. Rev. 24, 1781e1796.
Fedele, F.G., Giaccio, B., Hajdas, I., 2008. Timescales and cultural process at
40,000 BP in the light of the Campanian Ignimbrite eruption, western Eurasia.
J. Hum. Evol. 55, 834e857.
Giaccio, B., Hajdas, I., Peresani, M., Fedele, F.G., Isaia, R., 2006. The Campanian
Ignimbrite tephra and its relevance for the timing of the Middle to Upper
Palaeolithic shift. In: Conard, N.J. (Ed.), When Neanderthals and Modern
Humans Met. Kerns Verlag, Tübingen, pp. 343e375.
Giaccio, B., Isaia, R., Fedele, F.F., Di Canzio, E., Hoffecker, J.F., Ronchitelli, A.,
Sinitsyn, A.A., Anikovich, M.A., Lisitsyn, S.N., Popov, V.V., 2008. The Campanian
Ignimbrite and Codola tephra layers: two temporal/stratigraphic markers for
the Early Upper Palaeolithic in southern Italy and eastern Europe. J. Volcanol.
Geoth. Res. 177, 208e226.
Goldberg, P., Conard, N.J., in press. Geoarchaeology and paleoenvironments. In:
Conard, N.J., Bolus, M., Münzel, S.C. (Eds.), Geißenklösterle II: Fauna, Flora und
UmweltverhältnisseimMittel- und Jungpaläolithikum, Kerns Verlag, Tuebingen.
Hahn, J., 1988. Die Geißenklösterle-Höhle im Achtal bei Blaubeuren I. Fundhor-
izontbildung und Besiedlung im Mittelpalëolithikum und im Aurignacien.
Konrad Theiss Verlag, Stuttgart.
Hajdas, I., Taricco, C., Bonani, G., Beer, J., Bernasconi, S.M., Wacker, L., 2012.
Anomalous radiocarbon ages found in Campanian Ignimbrite deposit of the
Mediterranean deep-sea core ct85-5. Radiocarbon 53, 575e583.
Higham, T.F.G., 2011. European Middle and Upper Palaeolithic radiocarbon dates are
often older than they look: problems with previous dates and some remedies.
Antiquity 85, 235e249.
Higham, T.F.G., Bronk Ramsey, C., Karavanic, I., Smith, F.H., Trinkaus, E., 2006a.
Revised direct radiocarbon dating of the Vindija G1 upper Paleolithic Nean-
dertals. Proc. Natl. Acad. Sci. 103, 553e557.
Higham, T.F.G., Jacobi, R.M., Bronk Ramsey, C., 2006b. AMS radiocarbon dating of
ancient bone using ultrafiltration. Radiocarbon 48, 179e195.
Higham, T.F.G., Brock, F., Peresani, M., Broglio, A., Wood, R., Douka, K., 2009. Prob-
lems with radiocarbon dating the Middle to upper Palaeolithic transition in
Italy. Quatern. Sci. Rev. 28, 1257e1267.
Higham, T.F.G., Jacobi, R.M., Julien, M., David, F., Basell, L., Wood, R., Davies, S.W.G.,
Bronk Ramsey, C., 2010. The chronology of the Grotte du Renne (France) and
implications for the association of ornaments and human remains within the
Châtelperronian. Proc. Natl. Acad. Sci. 107, 20234e20239.
Higham, T.F.G., Jacobi, R.M., Bassell, L., Bronk Ramsey, C., Chiotti, L., Nespoulet, R.,
2011. Precision dating of the Palaeolithic: a new radiocarbon chronology for the
Abri Pataud (France), a key Aurignacian sequence. J. Hum. Evol. 61, 549e563.
Hoffmann, D.L., Beck, J.W., Richards, D.A., Smart, P.L., Singarayer, J.S., Ketchmark, T.,
Hawkesworth, C.J., 2010. Towards radiocarbon calibration beyond 28 ka using
speleothems from the Bahamas. Earth Planet. Sci. Lett. 289, 1e10.
Hughen, K., Lehman, S., Southon, J., Overpeck, J., Marchal, O., Herring, C., Turnbull, J.,
2004.14C activity and global carbon cycle changes over the past 50000 years.
Science 303, 202e207.
Hughen, K., Southon, J., Lehman, S., Bertrand, C., Turnbull, J., 2006. Marine-derived
14C calibration and activity record for the past 50,000 years updated from the
Cariaco Basin. Quatern. Sci. Rev. 25, 3216e3227.
Jacobi, R.M., Higham, T.F.G., Bronk Ramsey, C., 2006. AMS radiocarbon dating of
Middle and Upper Palaeolithic bone in the British Isles: improved reliability
using ultrafiltration. J. Quatern. Sci. 21, 557e573.
Jöris, O., Street, M., 2008. At the end of the14C time scale-the Middle to Upper
Paleolithic record of western Eurasia. J. Hum. Evol. 55, 782e802.
Langley, M.C., Clarkson, C., Ulm, S., 2008. Behavioural complexity in Eurasian
Neanderthal populations: a chronological examination of the archaeological
evidence. Cambridge Archaeol. J. 18, 289e307.
Longin, R., 1971. New method of collagen extraction for radiocarbon dating. Nature
230, 241e242.
Mellars, P., 1999. The Neanderthal problem continued. Curr. Anthropol. 40,
341e364.
Mellars, P., 2005. The impossible coincidence. A single-species model for the origins
of modern human behavior in Europe. Evol. Anthropol. 14, 12e27.
Münzel, S.C., Conard, N.J., 2004. Change and continuity in subsistence during the
Middle and upper Paleolithic in the Ach valley of Swabia (SW-Germany). Int. J.
Osteoarchaeol 14, 223e225.
Münzel, S.C., in press. Die Großsäugetierreste aus dem Geißenklösterle. In: Conard,
N.J., Bolus, M., Münzel, S.C. (Eds.), Geißenklösterle II: Fauna, Flora und Umwelt-
verhältnisse im Mittel- und Jungpaläolithikum, Kerns Verlag, Tuebingen.
Pyle, D.M., Ricketts, G.D., Margari, V., van Andel, T.H., Sinitsyn, A.A., Praslov, N.D.,
Lisitsyn, S., 2006. Wide dispersal and deposition of distal tephra during the
Pleistocene ‘Campanian Ignimbrite/ Y5’ eruption, Italy. Quatern. Sci. Rev. 25,
2713e2728.
Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Blackwell, P.G., Bronk
Ramsey, C., Buck, C.E., Burr, G., Edwards, R.L., Friedrich, M., Grootes, P.M.,
Guilderson, T.P., Hajdas, I., Heaton, T.J., Hogg, A.G., Hughen, K.A., Kaiser, K.F.,
Kromer, B., McCormac, F.G., Manning, S.W., Reimer, R.W., Richards, D.A.,
Southon, J.R., Talamo, S., Turney, C.S.M., van der Plicht, J., Weyhenmeyer, C.E.,
2009. Intcal09 and Marine09 radiocarbon age calibration curves, 0e50,000
years cal BP. Radiocarbon 51, 1111e1150.
Richter, D., Waiblinger, J., Rink, W.J., Wagner, G.A., 2000. Thermoluminescence,
electron spin resonance and14C-dating of the Late Middle and Early Upper
Palaeolithic site of Geißenklösterle Cave in southern Germany. J. Archaeol. Sci.
27, 71e89.
Staff, R.A., Bronk Ramsey, C., Nakagawa, T., Suigetsu 2006 Project Members, 2010.
A re-analysis of the Lake Suigetsu terrestrial radiocarbon dataset. Nucl. Instrum.
Meth. B 268, 960e965.
Svensson, A., Andersen, K.K., Bigler, M., Clausen, H.B., Dahl-Jensen, D., Davies, S.M.,
Johnsen, S.J., Muscheler, R., Rasmussen, S.O., Rothlisberger, R., Steffensen, J.P.,
Vinther, B.M., 2006. The Greenland ice core chronology 2005, 15e42 ka. Part 2:
comparison to other records. Quatern. Sci. Rev. 25, 3258e3267.
Svensson, A., Andersen, K.K., Bigler, M., Clausen, H.B., Dahl-Jensen, D., Davies, S.M.,
Johnsen, S.J., Muscheler, R., Parrenin, F., Rasmussen, S.O., Röthlisberger, R.,
Seierstad, I., Steffensen, J.P., Vinther, B.M., 2008. A 60,000 year Greenland
stratigraphic ice core chronology. Clim. Past 4, 47e58.
Szmidt, C.C., Normand, C., Burr, G., Hodgins, G., LaMotta, S., 2010. AMS14C dating
the Protoaurignacian/Early Aurignacian of Isturitz, France. Implications for
Neanderthal-modern human interaction and the timing of technical and
cultural innovations in Europe. J. Archaeol. Sci. 37, 758e768.
Talamo, S., Soressi, M., Roussel, M., Richards, M., Hublin, J.-J., 2012. A radiocarbon
chronology for the complete Middle to Upper Palaeolithic transitional sequence
of Les Cottés (France). J. Archaeol. Sci. 39, 175e183.
Teyssandier, N., Liolios, D., 2003. Defining the earliest Aurignacian in the Swabian
Alp: the relevance of the technological study of the Geißenklösterle (Baden-
Württemberg, Germany) lithic and organic productions. In: Zilhão, J., D’Errico, F.
(Eds.), 2003. The Chronology of the Aurignacian and of the Transitional Tech-
nocomplexes: Dating, Stratigraphies, Cultural Implications. Trabalhos de
Arqueologia, vol. 33. Instituto Português de Arqueologia, Lisbon, pp. 179e196.
Teyssandier, N., Bolus, M., Conard, N.J., 2006. The Early Aurignacian in central
Europe and its place in a European perspective. In: Bar-Yosef, O., Zilhão, J. (Eds.),
2006. Towards a Definition of the Aurignacian. Trabalhos de Arqueologia, vol.
45. Instituto Português de Arqueologia, Lisbon, pp. 241e256.
Voelker, A.H.L., Grootes, P.M., Nadeau, M.-J., Sarnthein, M., 2000. Radiocarbon levels
in the iceland sea from 25e53 kyr and their link to the earth’s magnetic field
intensity. Radiocarbon 42, 437e452.
Ward, G.K., Wilson, S.R., 1978. Procedures for comparing and combining radio-
carbon age determinations: a critique. Archaeometry 20, 19e31.
Weninger, B., Jöris, O., 2008. A14C age calibration curve for the last 60 ka: the
Greenland-Hulu U/Th timescale and its impact on understanding the Middle to
Upper Paleolithic. J. Hum. Evol. 55, 772e781.
Wood, R.E., Bronk Ramsey, C., Higham, T.F.G., 2010. Refining the ultrafiltration bone
pretreatment background for radiocarbon dating at ORAU. Radiocarbon 52,
600e611.
Wood, R.E., Higham, T.F.G., de Torres, T., Tisnérat-Laborde, N., Valladas, H., Ortiz, J.E.,
Lalueza-Fox, C., Sánchez-Moral, S., Cañaveras, J.C., Rosas, A., Santamaría, D., de la
Rasilla, M. A new date for the Neanderthals of El Sidrón Cave (Asturias,
Northern Spain). Archaeometry, in press.
Zilhão, J., 2006. Genes, fossils, and culture. An overview of the evidence for
Neandertal-modern human interaction and admixture. Proc. Prehist. Soc. 72,
1e20.
Zilhão, J., D’Errico, F., 1999. The chronology and taphonomy of the earliest Auri-
gnacian and its implications for the understanding of Neandertal extinction.
J. World Prehist 13, 1e68.
Zilhão, J., D’Errico, F., 2003a. The chronology of the Aurignacian and transitional
technocomplexes. Where do we stand?. In: Zilhão, J., D’Errico, F. (Eds.), 2003a.
The Chronology of the Aurignacian and of the Transitional Technocomplexes:
Dating, Stratigraphies, Cultural Implications. Trabalhos de Arquelogia, vol. 33
Instituto Português de Arqueologia, Lisbon, pp. 313e349.
Zilhão, J., D’Errico, F., 2003b. An Aurignacian Garden of Eden in southern Germany?
An alternative interpretation of the Geissenklösterle and a critique of the Kul-
turpumpe model. Paléo 15, 69e86.
T. Higham et al. / Journal of Human Evolution xxx (2012) 1e13
13
Please cite this article in press as: Higham, T., et al., Τesting models for the beginnings of the Aurignacian and the advent of figurative art and
music: The radiocarbon chronology of Geißenklösterle, Journal of Human Evolution (2012), doi:10.1016/j.jhevol.2012.03.003