¿ìè¶ÌÊÓƵ

About 14 000 years ago an ice-age hunter painted three extraordinary
bison on the ceiling of a cave. Depicted in red, black and yellow, with
gently folded legs and naturalistic forms, these bison, among many to be
found in the Altamira Cave in northern Spain, are a powerful testimony to
the creative talents of an ancient artist. Or so archaeologists thought.

But they were wrong. Last May a team of French and Spanish archaeologists
published results from new analysis of these works of art, showing them
to be painted not by a single artist but by several people at three different
times – 13 570, 13 940 and 14 330 years ago.

The evidence for these dates came from minute samples of charcoal in
pigment taken from the paintings. Only in the past ten years have researchers
discovered organic materials in some prehistoric pigments, and devised ways
of isolating them for analysis. And during that decade they have perfected
a technique to allow dating of such tiny samples of carbon. This technique,
accelerator mass spectrometry (AMS) radiocarbon dating, is increasingly
in demand. It is fast becoming established as the first of a new generation
of analytical tests enabling archaeologists to date ancient rock art directly
instead of relying solely on circumstantial evidence.

Like the traditional radiocarbon dating used in other branches of archaeology,
AMS estimates the age of a sample of organic material from the amount of
radioactive carbon-14 it contains . The older a sample, the more time its
carbon-14 has had to decay (half the carbon-14 is lost every 5730 years)
and the smaller the proportion of this isotope compared with the two stable
isotopes, carbon-12 and carbon-13. But AMS differs radically from conventional
radiocarbon dating in the way that carbon-14 is detected, which is why it
can be used to date minute carbon samples.

Normally, the amount of carbon-14 is assessed indirectly by measuring
the rate of its radioactive decay – around 3 grams of organic material
is required. In AMS the carbon sample is smashed up into its constituent
atoms, and the number of atoms of each isotope is then counted. A sample
as small as 100 micrograms can be successfully dated. This advantage is
crucial when it comes to dating ancient rock paintings like those in the
Altamira Cave, from which only a few milligrams of pigment may be scraped
without causing unacceptable damage.

Before AMS radiocarbon dating, the best that rock art researchers could
hope for was to find a petroglyph (rock carving) covered by a pile of ancient
rubbish, or a piece of a pictograph (rock painting) that had flaked off
a rock face and landed among the litter of prehistoric life. By obtaining
a radiocarbon date for deposits associated with the art, experts could set
a ‘minimum age’ for it, says Jack Steinbring, an archaeologist from the
University of Winnipeg. That is, they could be reasonably sure that the
art was at least as old as the artefacts associated with it.

Age could also be inferred using methods from fine-art history. Prehistoric
‘schools’ of painting have been identified by asking questions about style.
What kinds of shapes do the figures have? How are they executed: incised,
painted or pecked into the surface? This approach has enabled experts to
piece together chronologies of styles for various geographical areas. These
are then linked to archaeological remains dated using radiocarbon techniques.
But appearances can be deceptive, as the new findings from the Altamira
Cave prove.

Worn away by weather

Fortunately, style and an association with prehistoric detritus are
not the only indirect clues to a painting’s age. Weathering is another,
though not very accurate, indicator. Older works of art tend to be more
eroded, but this depends heavily on the severity and speed of changes in
the local environment over thousands of years. Researchers can also get
a feel for the age of a painting where the rock has been covered by a natural
deposit, by estimating how long it took for such processes to occur.

By applying these techniques – separately and together – experts believe
they have good dates for numerous rock art sites. But these dates remain
best guesses, the only way to know the true age of rock art is to test the
painting or carving itself.

Enter Erle Nelson and his team at Simon Fraser University in Canada.
As a nuclear physicist specialising in archaeometry back in the 1970s, Nelson
realised that if he could devise a method of radiocarbon dating capable
of analysing minute samples of organic material, this would greatly increase
its potential for use in the dating of rock art. After all, the traditional
technique gave reliable dates as far back as 40 000 to 45 000 years, easily
covering the time span when experts believe most prehistoric art works were
created. He hit on the idea of actually counting the number of carbon-14
atoms in a small sample, and AMS radiocarbon dating was born. With the technique
being simultaneously developed at the University of Toronto and at the University
of Rochester in New York, disputes have since developed about which team
invented it, but as Nelson concedes: ‘It was an idea whose time had come.’

AMS radiocarbon dating is only now coming into its own, however. At
first there was some resistance to its use by archaeologists, few of whom
have any training in nuclear physics or chemistry. But gradually they began
to realise that it could supplement traditional methods. Its popularity
has been paralleled by refinement of the technique to make it more reliable,
and new discoveries revealing the extent to which organic materials were
used in ancient pigments – not just charcoal, but natural products such
as blood, honey, milk and oil seed, used to bind the paints. Now there are
more than 20 centres for AMS dating worldwide.

Michel Lorblanchet, a researcher at the CNRS, France’s national research
centre, is one of the greatest advocates of AMS dating. He estimates that
fewer than 20 rock art sites around the world have been dated using the
technique. Given that millions of pictographs and petroglyphs have been
found on every continent except Antarctica – Australia’s Kakadu National
Park, alone, has more than 20 000 known sites – the progress seems remarkably
slow. One reason is that AMS dating is not cheap. New or refitted accelerators
cost from $3 million, and dating of individual samples ranges from about
$300 to $900. But a more fundamental obstacle is the enormous amount of
time and expertise needed to prepare samples.

Preparing for testing times

Running a sample through an AMS system takes less than an hour, but
this is merely the final step in a complex and time-consuming process. First,
there is the problem of obtaining a sample without harming the cultural
or artistic integrity of a work of art. Selection of samples is crucial
because they must come directly from the pictograph to be analysed. Contaminants
from subsequent natural geological, biological or environmental events –
such as ash from a forest fire or animal faeces – will distort the age.

Extraction of the carbon from a sample is also difficult. Most AMS laboratories
use an approach developed by John Vogel, a member of Nelson’s original team.
A sample is carefully washed in distilled water and placed in a quartz tube
along with hydrogen gas and a metal such as iron or cobalt, onto which the
carbon is deposited once separated from the sample. A small electric oven
heats the tube to about 600 degreesC, causing the hydrogen to bind with
oxygen in the pigment to form water which freezes at one end of the tube.
The remaining graphite deposit can either be used directly as the source
of carbon for AMS analysis or can be converted first to carbon dioxide.

This process is slow, but Marvin Rowe, a physicist and nuclear chemist
at Texas A & M University, and his colleague, chemist Marion Hyman,
are exploiting advances in plasma chemistry in experiments that may help
to speed up extraction. They place the sample in an airtight chamber containing
an oxygen plasma – a highly reactive mixture of positively charged atomic
nuclei in a sea of electrons stripped from them. The plasma reacts only
with the organic carbon in the sample and ‘burns’ it off as carbon dioxide.
This is then collected in a sealed tube as dry ice. The whole procedure
takes about a week, and initial tests, comparing the results with those
from previously dated samples, suggest it is accurate. But even if this
technique were adopted widely, it would do little to reduce the time – frequently
as much as a year – that archaeologists must wait for samples to be analysed.
There is simply far too much demand for the available AMS dating capacity.

Now the technique is even being used to give more accurate dates for
rock art that has no carbon directly associated with it – engravings, or
paintings with inorganic pigments, for example. This has been made possible
by developments in methods for collecting samples from the translucent crusts
that form over many pictographs and petroglyphs in desert conditions. This
natural varnish or ‘skin’ is deposited layer upon layer as the climate changes.
Individual analysis of these microscopically thin layers would allow researchers
to build up a ‘micro-stratigraphic context’, turning rock skins into historical
maps. The pioneer of this approach is Alan Watchman, a PhD student at the
Australian National University, now living in Quebec, Canada.

A rock skin is typically composed of clay minerals, oxides, hydroxides
of manganese and iron, trace elements and a minute amount of organic matter,
mainly in the form of oxalates. The most important oxalate, oxalic acid,
comes from urine sprayed onto rock walls by passing animals, from deposits
left by some lichens, algae and fungi, and from decaying vegetation. In
1987 Watchman first demonstrated that carbon from these oxalates could be
dated with AMS. His technique uses a laser to burn the organic material
from each layer in a localised area, giving carbon dioxide for analysis
by AMS radiocarbon dating. Focused Laser Extraction of Carbonaceous Substances
(FLECS), has already given accurate and reliable dates.

AMS radiocarbon dating has undoubtedly had a huge impact on the study
of prehistoric works of art. But Jean Clottes, archaeologist and director
of antiquities for the French Ministry of Culture, cautions colleagues who
embrace the new dating technology with too much fervour. ‘We archaeologists
have a naive fascination with hard science,’ he says. As a pioneer of AMS
dating, Nelson agrees that the technique has its limitations, but he also
views it in a wider context, as heralding a more scientific approach in
this field of research. ‘It’s going to take a lot of hard work and thought,’
he says, ‘but if we can do it – and I’m very optimistic – we’ll bring rock
art out of the never-never land of archaeology, right into the mainstream.’

* * *

Counting on a date with the past

Traditional radiocarbon dating measures the rate of radioactive decay
of carbon-14 to determine the amount of the isotope present. Living material
has the same ratio of radioactive carbon-14 to nonradioactive carbon-13
and carbon-12 as is found in the atmosphere. After death, however, the carbon-14
content decays, half being lost every 5730 years. So by comparing the proportions
of these three a sample can be dated.

Accelerator mass spectrometry (AMS) radiocarbon dating measures the
proportion of carbon-14 atoms directly, by counting the number of atoms
of each carbon isotope in a small sample of organic material. The technique
requires the coupling of a particle accelerator and a mass spectrometer
to purify and then separate the carbon fractions.

Comparison with results from conventional radiocarbon dating shows that
it is accurate. But the main advantage is that it can be used on minute
samples of organic material – as small as 100 micrograms, as opposed to
3 grams for the traditional method – allowing analysis of samples that
could not previously be dated.

First the carbon atoms are ionised into negative ions by the addition
of electrons. As they pass from the ion source into the accelerator they
are deflected through about 90degrees by a magnet. This procedure removes
any nitrogen-14, which would otherwise be detected along with the carbon-14
fraction. The speeding ions then pass through a thin foil or a gas which
strips off electrons to leave them positively charged. Any molecular contaminants
– carbon-13 with a hydrogen atom stuck to it, for example – are eliminated
in the process. Finally, the isotopes are separated according to mass by
accelerating them through one or more magnetic fields where they are deflected
down different beam lines to detectors. Here they are measured and counted.