AMS Radiocarbon Dating of a Wood Sample from Tunel Ravne:
Methods, Results and Implications for Further Research.
Andrew Lawler BA (Hons) Cantab
Abstract
This paper discusses the discovery of a piece of wood discovered embedded in the sediments of the Tunel
Ravne complex, Visoko, Bosnia-Herzegovina. The wood was sampled, and sent for Radiocarbon analysis
to two independent establishments. Below the procedures involved in sampling, and a brief outline of the
methodologies and principles applied in the AMS 14C dating are detailed, as well as an appraisal of possible
implications for further work and suggested directions of future research regarding the tunnel system.
Introduction
During work on Tunel Ravne November 2007, workers discovered a dark, soft
compressible material preserved embedded within the sides of the tunnel, at a height of
1.4 metres, during the clearance of debris and installation of wooden support beams to
facilitate visitor access. Initial examination on 12th November by employees of the
Foundation led to the assumption that the material was organic in nature, and potentially
capable of yielding dating evidence. The Permanent Archaeologist was alerted to this,
and work in the immediate vicinity was temporarily halted.
Originally, the material measured approximately 35cm in length and 8cm in width, with
an unknown depth (now estimated to have been approximately 12cm), but a small
collapse prior to 10th December reduced its width by approximately 8cm.
On 15th January 2008, the Permanent Archaeologist took a small sub-sample of the
material, and confirmed it to be wood. After discussion with directors of the Foundation,
it was decided that this wood should be sampled, and C-14 analysis of the samples should
be undertaken by
independent
institutions, in order
to determine its age.
Figure 1- The wood in
situ, with label
1
Sampling
On 29th January 2008, the Permanent Archaeologist, along with an assistant,
photographed the wood in situ, and removed 4 samples (<5a>, <5b>, <5c>, <5d>), each
weighing approximately 5 grams. These were placed in labelled polythene bags, in
accordance with guidance given in a telephone conversation with a member of staff at the
Research Laboratory for Archaeology and the History of Art, University of Oxford.
Figure 2- The 4 Samples
After removal, a label was embedded into the
sediments, denoting the location and date of sample
removal, alongside the index number. The samples
were then taken and stored in a refrigerator at 4°c,
whilst the institutions to carry out the analyses were
selected.
Eventually, it was decided that two separate AMS
laboratories were to perform the analysis. Those chosen were the Research Laboratory
for Archaeology and the History of Art, University of Oxford, UK, and the LeibnizLaboratory for Radiometric Dating and Stable Isotope Research, of Christian-Albrechts
University, Kiel, Germany. After completing all necessary paperwork and data sheets,
these samples were packed in protective envelopes, and sent by recorded delivery on 26th
February, 2008.
Figure 3- Oxford sample and data sheets ready to be sent
2
Notification that the samples were delivered to the relevant institutions was received by
the Foundation on the 5th and 7th of March from Oxford and Kiel respectively.
Results- Kiel
On 6th May 2008, the Foundation received the results of the analysis undertaken by the
Leibniz-Laboratory for Radiometric Dating and Stable Isotope Research, at ChristianAlbrechts University, Kiel, via email. Below is a summary of them:
Table 1- Summary of 14C AMS Results
Fraction
Corrected pMC†
Wood, alkali residue, 1.2 mg C
2.22 ± 0.14
Conventional Age
13C(‰)‡
30600 + 540 / -510 BP
-26.47 ± 0.08
Preparation, Methodology and Calculation
After receipt, the wood sample was checked under a microscope and all material suitable
for dating was selected. The demineralisation process consisted of processing through
solutions of 1 % HCl, 1 % NaOH at 60°C and again 1 % HCl (to eradicate any alkaline
residue). The combustion of the demineralised material into CO2 was performed in a
closed quartz tube together with CuO and silver wool at 900 °C. The sample CO2 was
reduced with H2 over about 2 mg of Fe powder as catalyst, and the resulting carbon/iron
mixture was pressed into a pellet in the target holder, ready for AMS analysis. Alongside
sample 5C, samples of the international isotopic standards used in AMS 14C dating, being
reduced CO2 derived from oxalic acid (for ‘present day’ levels) and from background
coal (considered to be radiocarbon ‘dead’) were also included in the particular AMS
‘run’, for comparison.
Equipment
The AMS system employed by the Leibniz-Laboratory for Radiometric Dating and
Stable Isotope Research is a HVE 3 Million Volt Tandetron 4130 AMS system, equipped
with a single caesium sputter ion source.
3
Methodology
The AMS methodology achieves accurate and reliable 14C dates by measuring the ratios
of different carbon isotopes through the use of a high-powered electromagnet to separate
single ionised atoms in a gaseous state according to atomic mass. These are deflected by
the electromagnets into separate ‘chambers’, where their high-velocity impacts are
counted individually. The fact that atom impacts are individually counted allows the
accurate measurement of carbon isotope ratios from relatively low amounts of organic
material (around 20mg is the generally accepted minimal amount). In the case of
14
C
AMS dating, the isotopes studied are those of all three carbon isotopes (12C, 13C, 14C). In
nature, Nitrogen with an atomic mass of 14 also occurs naturally (as well as several
13
and
12
C-
C- bearing molecular ions). To eradicate the risk of inaccuracies through the
accidental measurement of 14N (which would give a much younger date, as particles with
an atomic mass of 14 would be much more abundant in results), the gaseous samples are
initially run in a negative ion beam, as N- ions are not stable. The AMS machine’s
counters are then set to measure atomic masses of 12, 13, and 14, by adjusting the
strengths of the electromagnets, and the individual ions are propelled into their relevant
counters.
After the ions have been propelled into the counters and counted, the ratio of the three
carbon isotopes is then calculated in order to calculate the age of the material studied.
Implications of result
During personal discussions with Mr Semir Osmanagić and Mr Goran Čakić, several
statements were made regarding the circumstances in which the wood was found,
namely:
-
The wood was totally encased within the conglomerate material.
-
The conglomerate material (removed in 2006) that overlay much of the nearby K1
megalith was in its original, compacted state.
-
Many of the carvings evident on this stone were concealed by the consolidated
conglomerate.
4
Supposing that the wood was deposited by the geological event which created the
sediments into which the tunnels are cut, then the two possible conventional date ranges
given by the Kiel AMS laboratory are in agreement with the proposed age of sediments
given in an assessment report by Dr Aly A. Barakat after his visit to Visoko in 2006, in
which he states that the Ravne continental conglomerate deposits are geologically recent,
and formed during the Pleistocene or Holocene epoch. Regarding the above points, this
material would be able to provide a terminus ante quem for the carvings observed on the
K1 megalith, unless another event had reconsolidated collapsed material after the creation
of the tunnel system, such as localised flooding or calcitic leaching. Unfortunately, scant
documentation of this material in-situ and its subsequent removal exists, and no samples
were taken of it for comparative analyses by independent institutions. Therefore, it cannot
be proved that these carvings predate the original deposition of these sediments. More
likely is the scenario in which these megaliths were deposited as erratics by the
geological processes involved in the formation of the Ravne sediments. After these
deposits stabilised and natural concretion processes evolved them into conglomerates,
they became a suitable medium for tunnelling, and the tunnels were created, at a date as
yet unknown. As these erratics were encountered, carvings were made on them. The
tunnels were then abandoned at a later date, minor collapses of the conglomerate material
occurred, and the erratics were once again buried under loose sediments. These were then
able to reconsolidate to form a secondary conglomerate, derived from the original
material.
At some points throughout the tunnel system, grading (see Figure 4) of material within
the conglomerate can be observed, suggestive of original deposition within an aquatic
environment. The fact that the carvings were cut into sandstone, a material highly
susceptible to erosion and weathering, suggests that they were carved at a date long after
these depositional processes had finished, as it is unlikely that they would survive for
long in a high-energy riverine or coastal environment. This further backs up the idea that
the stones and their carvings were reburied under collapsed material after the creation of
the tunnels.
5
Figure 4- Grading evident in Tunnel Sediments (15m from Sample)
Large-scale geological maps of the area created in 1967 and 2002 do not mention the
deposits identified by Dr. Barakat, but instead depict the geology in the vicinity of the
Ravne tunnel system as dating from the Upper Miocene (11.6-5.3 mya) period. However,
the footnotes of the latter of the two maps do acknowledge the presence of minor
geological unconformities (Cicić, 2002). These maps can help in providing the basis for
alternative theories for the deposition of the wood in the Ravne system. It is plausible that
the tunnel system is, in fact, cut into the aforementioned Upper Miocene deposits, which
in turn leads to a wider range of possibilities regarding the method of deposition and
covering of the wood, providing a different set of archaeological implications.
Firstly, it is possible that the tunnels were cut into the Upper Miocene deposits prior to
the deposition of the Pleistocene or Holocene deposits identified by Dr. Barakat in his
report. These may have then been deposited by events in the Quaternary Period such as
localised flooding from external sources, most likely a river or glacial meltwater,
resulting in the creation of a layer of later conglomerates, or even several layers built up
over periods of flooding within what was originally a much larger space, an idea which
6
could be supported by the grading evident in Figure 4, and the fact that the concretion
and cohesivity of the sediments vary to quite a large extent throughout the tunnel
network. The idea that the sediments evident as the walls of the tunnel today formed as a
result of events after the initial creation of the tunnels leads to two scenarios by which the
dated wood could have been deposited. One is that it came into the tunnel during an
incursion of floodwater responsible for the deposition of other sedimentary material
during the Pleistocene or Holocene epoch, and buried by this material, which
reconsolidated to form the conglomerates visible today as the side walls of the tunnel
network in its current state of excavations. Alternatively, the wood could have been
embedded into the tunnel wall as a result of human activity, and later covered and
preserved within sediments deposited by floodwaters at a later date in the Quaternary
Period. Possible reasons for human placement of this wood are difficult to determine. The
apparent grain of the wood (see Figure 1) in its position when discovered is not
conducive to its use as a structural support. Admittedly, however, what appeared to be the
wood’s grain may not be such, but instead a fracturation pattern within the waterlogged
wood opposing the original grain, due to the high pressures and stresses imposed upon it
over time whilst encased within the re-deposited material.
Implications for Further Research
Due to the fact that the above inferred scenarios are dependant on two separate sets of
geological interpretations, with hugely differing archaeological implications, the priority
in regards to further research leading to an accurate dateline of events in the Ravne tunnel
system should be to ascertain the geological age of the medium into which the tunnels
were originally cut. To gain a fuller understanding of the accumulation and consolidation
of the conglomerates evident today as the current walls of the tunnel system, further
geological research is desirable, including an independent study of the conglomerate
matrix itself, culminating in the creation of a localised geological map, and also possibly
several cross-sectional maps of the sediments apparent as the tunnel walls, created
through the application and study of core sections. This method will aid in our
understanding of the sequence of archaeological and geological processes which occurred
7
after the initial creation of the Ravne tunnel system. However, as this will be a laborious
and time consuming process, involving a variety of intrusive and destructive methods, the
author recommends that, as the date ranges for the wood given the AMS 14C laboratories
comply with Dr. Barakat’s suggested age of the sediments, an emphasis of immediate
future research within the Ravne tunnel system should be placed upon the geological
processes which occurred after the abandonment of the complex. Such an approach is less
destructive, as the emphasis of research is placed upon the uppermost (most recently
deposited) sediments, which are routinely removed in the excavation and widening for
public access of the tunnel system.
Samples of material overlying stones which bear carvings that are uncovered in the future
should be maintained for both comparative geological analysis, and microscopic analysis,
in order to retrieve any microfaunal and floral remains, as well as possible organic
material, which could yield dating evidence. This methodology not only provides a
further insight into the processes which occurred after the abandonment of the tunnels,
but the dating evidence which could be collected in such an excavational procedure could
provide a terminus ante quem for both the carvings found upon future stones, and for the
tunnel system as a whole.
It should be noted that a further understanding of the chronological relationship between
past archaeological and geological events can also potentially be achieved through a
study of the relationship between the compacted sedimentary material and side walls
which frequently occur throughout the tunnels. As of this writing, one excavation with
such a purpose is currently being undertaken approximately 48 metres from the tunnel
entrance. This excavation is scheduled to be completed by the end of the 2008 season.
Note
As of July 2008, the result of the Research Laboratory for Archaeology and the History
of Art, University of Oxford’s is still awaited by the Foundation, due to delays in
analyses caused by the recalibration of the department’s accelerator mass spectrometer in
April 2008.
8
References
Earth Sciences Institute of Sarajevo. 2002. Geological map of Bosnia and Herzegovina at
1:300 000.
With Cicić, S. 2002. (Transl. Šahinović, D. 2008.) Accompanying Text.
Geology and Hydrogeology Institute, Civil Engineering Faculty, Sarajevo. 1964.
Geological map at 1:100 000 of the Visoko Region.
Sternberg, RA, Damon, PE. 1992. Implications of dipole moment secular variation from
50,000-10,000 years for the radiocarbon record. Radiocarbon 34(2):189-198.
Stein, M., Goldstein, S.L. & Schramm, A. 2000. Radiocarbon calibration beyond the
dendrochronology range. Radiocarbon 42, 415-422
Stuiver, M. & Polach, H.A., 1977. Discussion: Reporting of 14C Data. Radiocarbon
19(3): 355-363
Barakat, A.A., 2006. Geological and Geoarcheaological Observations on the Bosnian
Pyramids at Visoko, Northwest Sarajevo.
http://www.c14dating.com/agecalc.html
http://physicsworld.com/cws/article/news/2676
http://www.earth.ox.ac.uk/research/2007/CMN1.pdf
9