?

/f)/5/CS Z< AR<;HAEOLOGY

!i2fu,K,/ . Z/v7`1Q0ju 770/V

(Hb/0 LARBON .D/I WN
and 07//ER K/LO/0/VE77?/;
Tzcw/vx&us

L'fu,_~gg . L UM//VESCE/V c JH T/M

,7 W/RMOz. U/V//VESCE/VCE
`/U OPT/CAL .D/Q7?/\A
/;i/Qi; /76'GA/ff/SM M;
ARCH/}`OL0y OCR Output
Cos

77 + IFN
/? barn;

7* @/1
/4 /4.

5?30

/3
/Z83
EGM/L /BR/U AMOUA/7*
.. $-5
_ X " bbn
OOO,

`. CONCENTRHT/O WHEN

z
/0

5 /5 mon
A
M ef C OCR Output
 OCR OutputOCR OutputOCR OutputOCR OutputOCR OutputOCR OutputOCR OutputOCR OutputOCR OutputOCR OutputOCR OutputOCR OutputOCR OutputOCR OutputCERN/1__GEl;lERAL + ommo i3r`Cl`ii i. less 1

CERN Lecture 1: Mar 20,'95: Physics & Archaeology: /\/OI.-ES

(t = transparency) (s = slide)

Physics and Archaeology now forms part of the wider field usually
described as Archaeological Science and sometimes as Archaeometry --- a
Greek-based word meaning measurement of old things. This wider field
incorporates astronomy, chemistry, geology, and biology, together with all
their sub disciplines.

From botany one has the powerful but conceptuably simple dating technique
of treering counting which now takes us back to the end of the last lce Age
some 10,000 years ago. Not only are the rings a basis for dating but they
also provide a climatic record, with linking-in to major volcanic eruptions
for good measure.

From organic chemistry one has analysis of food residues in cooking vessels
and hence information on ancient diet. This too is provided by isotope
\_. analysis of bones, which can for instance tell us whether the main source of
food was from the sea or from the land.

From geochemistry one has confirmation that the monoliths of Stonehenge

did indeed come from South Wales but, taking into account late eighteenth
century geological observations concerning the presence in the vicinity of
glacial eratics, they were transported by glacial action rather than the
gangs of slave labour that had rather fancifully been suggested and depicted.

One could go on -- but turning now to Physics & Archaeology one first of
all remarks that this section of Archaeological Science was first in the
field. ln the nineteenth century an italian by the name of Folgerhaiter
measured the remanent magnetization in pottery from Arezzo, concluding,
erroneously, that there had been a reversal of the earths magnetic field a
few centuries B.C.; although handicapped and mislead by artefacts of his
___ measuring apparatus this example does incidentally serve to illustrate that
the benefit is potentially in both directions and in later applications of
remanent magnetization archaeologically-dated pottery and pottery kilns
has provided detailed information about the past behaviour of the earth's
magnetic field, and hence ultimately contributing to an understanding of the
mechanism of its generation, that is otherwise inaccessible. Work on
remanent magnetization was continued by Emile Thellier in Paris during the
1940*s and 50's and subsequently in Oxford and elsewhere, including Geneva.
Magnetism and Archaeology will be the main topic of my third lecture,
which will also include mention of the immediately archaeologically-useful
techniques of geophysical surveying for location of buried remains.

l think l can claim with all due modesty that the present shape and size of
Physics and Archaeology owes much to the establishment of the Oxford
Research Laboratory for Archaeology, of which I was Deputy Director for
some 30 years. This laboratory was established through the joint efforts of
the late Viscount Cherwell and the late Professor Christopher Hawkes OCR Output
 s1] Cherwell & Hawkes
Cherwell, or Professor Lindemann (Churchills advisor, " the prof."), being in
charge of the Clarendon Laboratory, represented Physics, and Hawkes
represented Archaeology. They met at dinner at Christ Church one night in
the early 50s and in response to Cherwell's usual attack: "You Arts people
dont know what youre doing; why don't you use Science?" Hawkes
unexpectedly replied "l quite agree. What can we do about it" That was the
moment of conception and after a gestation period of 5 years the laboratory
was born. Earlier than that E.T. Hall, its Director for the first 30 years or
so, had been a research student under Cherwells supervision, developing X
Ray Fluorescent Analysis as a nondestructive technique. An early
application was to the so-called Piltdown Nlan, found in a quarry in southern
England
s2} Piltdown
that had been claimed as the missing link' between ape and man. But Hall's
finding that the staining was potassium permanganate, rather than naturally
due to leaching-in of iron oxide from the gravel burial surround, added
substantial weight to the conclusion that it was an elaborate hoax.

Of course in those days X-Ray Fluorescence had not been claimed by

chemistry as a routine technique and this was an impressive demonstration
of what physics could do for archaeology -- clise the Augean stables, an
activity subsequently amplified by thermoluminescence as I shall be
mentioning to-mprrow. More run of the mill, but contributing positively
rather than dismissively, was an application that Stuart Young undertook
using Hall's apparatus; this was to Chinese Blue and White vases, the
requrement of non-destructiveness being self-evident
s3] Blue & White
The question at issue was the source of the cobalt used for the blue glaze.
The analyses showed that before AD 1400 the cobalt was free from
manganese in contrast to later pieces when there were variable amounts.
s4] Results
This corresponded to the use of local Chinese ore, free from manganese,
until contact with the West, around AD 1400, brought in cobalt from a
variety of other sources.

Subsequently there has been and still is extensive use of essentially

physicsbased instruments in unravelling details of past pratices,
particularly trading routes and patterns; besides using minor and trace
impurities there has been powerful use of isotopic composition, for
instance for establishing the provenance of metal objects, notably in the
Aegean region.

But it is in dating that physics has been of absolutely critical importance in

archaeology, specifically prehistoric archaeology. The earliest calendar is
the Egyptian one, beginning with the First Dynasty, and the first pyramid,
around 3000 BC. s5] Pyramid
The pyramid illustrated is in fact a much later one, the early ones not being
photogenic. OCR Output
 <@l NER_&+_DATlNG 2w;i i. ima 3

Until the advent of physics in the form of radiocarbon dating the whole
chronology of the Old World was based on this Egyptian Calendar, which
consisted of lists of kings, fixed in sidereal years through the recorded
observation, some 1000 years after its start, of a datable astronomical
event (occurring in the ninth year of the reign of Sesostris lll). Before 3000
BC all dates were conjectural and the immediate impact of radiocarbon on
the conjectural dates for Neolithic was dramatic, upsetting previous ideas
about the pace of mans development - what had been thought to have
taken one or two millennia to acheive was now seen to have taken four or
five; in particular the beginning of Jericho, conjectured as at around 4000
BC. was pushed back to 7-8000 BC.

The idea of radiocarbon dating arose in the course of study of the effects of
cosmic rays on the earths atmosphere. This was by W.F. Libby and his group
at the University of Chicago in the late 1940s and he was subsequently
awarded a Nobel prize for development of the dating technique. This slide
$6] Animal
fancifully depicts the basis. Neutrons produced by cosmic-rays incident on
the upper atmosphere interact with atmospheric nitrogen to produce
carbon-14; this is the dominant way in which neutrons interact with
nitrogen or oxygen. The carbon-14 is incorporated into 'heavy carbon dioxide
which then circulates with the non-radioactive carbon dioxide of the
atmosphere. Atmospheric mixing is rapid giving uniform concentration
worldwide despite the pole/equator ratio of cosmic-ray neutron intensity
of 5/1. There is entry into plant life through photosynthesis and hence into
animal life; there is also mixing-in with the carbonate of the oceans.

The concentration level in these carbon exchange reservoirs' is a balance

between production and radioactive decay (halflife 5730 years) and to the
extent that cosmic-ray intensity remained constant in the past --- to
within about 10% --- the concentration level has been constant. But of
., course if there is cessation of exchange with the reservoir e.g. by formation
of cellulose molecules in wood or by incorporation into bone the level
decreases according to the 5730-year halflife and this is the basis of
dahng.

ln g the feasibility of dating the first question that Libby had to

answer was whether there was enough carbon-14 to measure, even in
modern wood, let alone wood formed several halflives ago. From balloon
measurements of cosmic-ray neutron production and from the crossection
for the reaction (which he measured as1.7 barns) he predicted a production
rate of 7.5 kg/year and hence for the halflife of 5730 years (1%/83 years)
the equilibrium amount of carbon14 on earth works out at 62 tons.

This is distributed throughout the carbon exchange reservoir (biosphere &

oceans) and from the size of this the expected ratio of carbon-14 to carbon
12 works out at 1.5 million million. equivalent to a decay rate of 15 per
minute per gram of carbon. Each decay is accompanied by emission of a beta OCR Output
 cEraN/i GENERAL ggtiuc imhi i. ms 4
particle (of maximum energy 160 keV) and Libby concluded that this made
the technique feasible even with the not-toosensitive Geiger counter that
he had available at that time.

The first test of his calculations was made on methane gas collected from
the City of Baltimore sewage plant. That being successful he went on to
measure wood of known date from tombs in Egyptian pyramids, again with
success. $7] Egyptian results
The way was then open for application, with the dramatic impact on
prehistory that l have just mentioned. lt was now prehistory according to
physics.

lt was not all plain sailing however. During the 60*s it became increasingly
evident that radiocarbon dates prior to 2000 BC tended to be a bit too
recent. The evidence for this was from two sources. Firstly, with improved
techniques (gas proportional counters) the ages for early Egyptian tombs
were a few hundred years to young. Secondly the same tendency to
underestimate age was evident when measurements were made on wood
dated by dendrochronology (tree-ring dating). Extension of treering dating
back to several thousand years BC is possible because of the technique of
matching the inner part of one tree with the outer rings of an earlier tree
$8] Tree-ring matchin
Fossil redwoods and bristlecone pines from California and Arizona, which
have life spans of one or two thousand years were used, and subsequently
European oak, so that dated wood of the last 10,000 years is now available
and this is used to correct raw radiocarbon dates. This slide shows one of
the early calibration curves so obtained $9] Calibration curve
The favoured explanation is in terms of reduced magnetic shielding of
cosmic-rays due to a lowering of the earths magnetic field. Reduced
shielding results in a higher production rate and hence a higher than
assumed carbon-14 level at t=0
t1]l\/lagnetic shieldin
Calibration beyond the range for which tree-ring dated wood is available is
by comparison with dates on coral dated by the uraniumseries
disequilbrium technique; an underestimate of around 3000 years is
indicated at 30,000 years ago. Beyond that there are predictions based on
determinations of the past intensity of the earth's magnetic field; these
suggest that at around 40,000 years ago the underestimate was near zero.

40,000 years is 7 half-lives, so the carbon14 activity is down by a factor

of over 100 on that at t=0; long counting times are necessary but the real
limitation is modern contamination, either intrinsically in the sample or
acquired during sample preparation. 1% of modern carbon in a sample of
infinite age produces an apparent age of 40,000 years.

With special facilities, including isotopic enrichment, 70,000 years can be

reached, but rountinely the limit is around 40-50,000 years.

During the 1980s accelerator mass spectrometry (AMS) began to be used, OCR Output
 cEr<i~ix1 GENERAL + oATlNo wiki;. i-gas 5
$10] AMS schematic
t2] Gas counting vs AMS
Main advantages over beta counting are enhanced throughput and smaller
sample size; the limitation of contamination remains but because of the
smaller sample size required intrinsic contamination can be tackled; in the
case of bone by using chromatography to select reliable amino acids e.g.
hydroxyproline.

The small sample size also gives access to a wide range of particularly
interesting problems of which the Turin Shroud is one s11] Turin Shroud
$12] Shroud result

Because of the 40-50,000 year limit in radiocarbon dating archaeologists

tend to ask whether there is another cosmogenic isotope that would reach
further back. Some possibilities that have been considered are:
t3] Other cosmogenic isotopes
Although the haIflives may be attractive these all fall down in respect of
geochemistry as far as archaeology is concerned; beryllium-10 and
aluminium-26 are possible candidates for sediment cores. Calcium41 is
the only one being seriously considered for archaeology because of its
occurrence in bone. lt is formed by the action of cosmdlc-ray neutrons on
calcium-40 in the top metre or so of soil and finds its way into animals via
uptake in vegetational roots. But quite apart from the severe problem of
detection sensitivity (concentration ratio in modern bone is down by a
factor of 100 on carbon-14 in 'modern' carbon) there is no worldwide
exchange reservoir and consequently the concentration ratio will be
dependent on soil conditions etc., with no way of knowing what the ratio in
the sample was at t=0.
It seems that carbon-14 is unique in having a remarkable combination of
characteristics:
t4] Carbon-14 characteristics

___ But there is now an array of non-cosmogenic techniques, some of which are
shown here:
t5] Ranges of technigues
One should also mention fission-track counting --- based on the
spontaneous fission of uranium-238; also potassium-argon dating, based on
the accumulation of argon produced in the radioactive decay of potassium
40.
t6] Potassiumargon
ln the case of uranium-series the build-up is of thorium-230:
t7]Uraniumseries
Luminescence will be the subject of to-morrow's lecture, together with
mention of electron spin resonance (ESR) and the involvement of magnetism
in archaeology will be discussed on Wednesday.

ln concluding a general survey of dating techniques one should mention also

two that are based on chemical change: hydration thickness of obsidian
artefacts: racemization of amino acids from Lform to D-form. One must OCR Output
 carmxi _eeNi;i=zAt + onine i-gee 6
add the proviso that being based on chemical change they are highly
dependent on environmental conditions, particularly temperature.

Finally, there is the remarkable development of the last two decades, still
actively proceeding, with something new almost in every other issue of
Nature last year. l refer to the astronomical dating of worldwide climatic
variations such as manifested most dramatically in the lce Ages. The
astronomical theory of climate was first developed by Milankovitch more
than 50 years ago; it refers to variation in energy received from the sun due
to small perturbations in parameters of the earth's orbit, these
perturbations resulting from the changing configuration of the planets:
t8] Parameters of earth's orbit
The energy variation is very slight and subtle mechanisms have to be
postulated in order to explain the correlation that is now observed. Some
climatic indicators are:
t9] Manifestations of long-term climate variabilit
Having established the correlation on the basis of the dating techniques
discussed earlier the procee is then turned around so as to provide more
accurate dates for the indicated variations. This might seem a long way
from archaeology but through the climatic indicators on archaeological
sites, with correlation to oxygen isotope variations, it is becoming possible
to transfer the astronomically-based timescale onto Palaeolithic sites.
The types and amounts of tree and vegetational pollen are the primary
indicator, and this pollen profile (PBO being a temperature indicator based
on a variety of pollen types) matches well with the marine isotope
variations:
tl O] Les Echets
The marine isotope variations are numbered from the top down; warm
stages have odd numbers and cold stages even numbers.
Another climatic indicator in long cores of sediment turns out to be
magnetic susceptibility (so l can justify this last transparency at any rate
as being appropriate to physiw-based dating):
tl l ] Lac du Bouchet
Here the isotope ratio is from a deep Artic ice core, more than 2 km long.
Sometimes there is the possibility of oxygen isotope variations on
archaeological, or archaeologically-related, sites and one can then match
into the ice cores very directly. Two examples from Switzerland:
tl 2] Dattnau and Gersensee

So now we have the opportunity to test the accuracy of radiometric

techniques against astronomicaally-calculated predictions; it is reassuring
to find that by and large they are standing up very well. OCR Output
 CERN/1 GENERAL + onine i.mia was 7
Bibliography and References
Aitken lvl.J. (1990) Science-based dating in archaeology . Longman, London
and New York, 274 pp.
Bassino F.C., Labeyrie L.D., Vincent E., Quidelleur X., Shackleton N.J. &
Lancelot Y. (1994) The astronomical theory of climate and the age of
the Brunhes-Nlatuyama magnetic reversal.Earth and Planetary Science
Letters 126. 91-108.
Guiot J., Pons A., de Beaulieu J.L. & Reille M. (1989) A 140,000-year
continental climate reconstruction from two European pollen records.
Nature 338. 309-313.
Hays ].D. Imbrie j., & Shackleton N .] (1976) Variations in the earth's orbit:
pacemaker of the ice ages. Science 194. 1121-1132.
Kaiser l<.F. (1993) Beitrage zur Klimageschichte vom spaten Hochglazial bis
ins fruhe Holozan rekonstruiert mit Jahrringen und Molluskenschalen aus
verschiederien l/ereisungsgebieten. Zeigler Druck& VerlagsAG,
Winterthur 203 pp.
Libby W.F. (1965) Radiocarbori Dating (2nd edn) University of Chicago Press,
Chicago.
Taylor, R.E. (1995) Radiocarbon Dating. ln Chronometric and Allied Dating in
Archaeology (Taylor, R.E. & Aitken, I\ll.J., eds.) Plenum Press, New York.
Thouveny N., de Beaulieu J.-L., Bonifay E., Creer l<.M., Guiot J., lcole lvl.,
Johnsen S., Jouzel J., Reille M., Williams T. & Williamson D. (1994)
Climate variations in Europe over the past 140 kyr deduced from rock
magnetism. Nature 371. 503-506. OCR Output
 CERN/1 GENERAL + DATING ~imia, less 8
CERN Lecture 1: Mar 20.'95 : Physics & Archaeology: TRANSP. & SLIDES
(t = transparency) (s = slide)

Lec/rw; 5 $1] Cherwell & Hawkes

$2} Piltdown

$3] Blue & White

$4] Results

$5] Pyramid

S6] @@1
_ , tg] ;mlbII|~h1 ` I Q ~ _ A ___

s7]Eg;gtiari results
$8] Tree-ring matchin

$9] Calibration curve

t1]Magnetic shieldin

$10] AMS schematic

t2] Gas counting vs AMS

s11]`l'urin Shroud

$12] Shroud result OCR Output

t3] Other cosmogenic isotopes

t4] Carbon1 4 characteristics

t5] Ranges of technigues

t6] Potassium-argon

t7] Uranium-series

t8] Parameters of earth's orbit

t9]Manife$tations of long-term climate variabilit

t10] Les Echets

t11] Lac du Bouchet

t12] Dattnau and Gersensee

 ctraiixz to 1
CERN Lecture 2 Mar21,'95 LUIVIINESCENCE DATING: Notes
(t = transparency) (s = slide)

ln the late 'I950's thermoluminescence (TL) was being developed for

radiation dosimetry (as an alternative to photographic film) and one of the
leading researchers, Farrington Daniels at Wisconsin, suggested that TL
might also be useful for dating in geology and archaeology.
It was tried in application to pottery about 35 years ago both by Kennedy in
Los Angeles and by the late Norbert Grogler in Berne (Grogler, Houtermans &
Stauffer 1960: He/vetia Physica Acta 33, 595-6); this was at the physics
institute there where Houtermans had been developing TL for use in
meteorite studies.
From a ground-up sample of pottery Grogler obtained a glow curve such as
tll Simple cllow-curve
Neither Kennedy nor Grogler pursued the idea of pottery dating and a few
years later we took it up at the Oxford Research Lab for Archaeology having
in mind its widespread application to archaeological pottery fragments,
undatable by radiocarbon, and art ceramics such as this Jernaa Head from
Nol<, W.Africa, which Stuart Fleming subsequently showed to be some 2500
years old s1l NoI< head
However before we reached that stage we had to overcome some major
difficulties:

The worst of these was that we didnt seem to get much dependence on age
and I remember the day of agonising reappraisal when gloomily discussing
things at coffee time we smashed a coffee cup and subsequently measured
its TL. Our worst fears were realised -- it gave as much TL as a Roman
pottery fragment. Something must be wrong! Eventually we found that it
was absolutely essential to do the heating in an atmosphere of dry nitrogen,
otherwise the 'true TL was swamped by parasitic 'spurious luminescence
from grain surfaces. The basic apparatus used
s2i ADDaratus
Of course it is now fully automated and computerised, with developments
such as CCD cameras being incorporated.
Luminescence of minerals has been studied for some time, being used for
mineral identification in the nineteenth century, and then for solid state
research from the beginning of the present century. But it was being
studied as long ago as 300 years as indicated by this first page
s3i Bovle
of a paper read to the Royal Society by Robert Boyle in October 1663. In it
he gave an account of observations he had made the previous night 'About a
Diamond that Shines in the Dark'. In accompanying notes he comments that OCR Output
 CERN/2 LD d
he wrote the paper during a drowsy period of the night when he wasnt fit
for making observations and that he brought the paper straight on with him
no nonsense about referees in those days! He also reported 'l also
brought it to some kind of Glimmering Light, by taking it into Bed [a 4
poster of course] with me, and holding it a good while upon a warm part of
my Naked Body. Elsewhere he comments that it was an extremely sensitive
diamond because 'as you know, my constitution is not of the hottest'.

Well, to return to the present, the essential basis of TL dating, as of optical

dating and ESR, is the build-up of trapped electrons resulting from
ionization caused by the weak flux of nuclear radiation from thorium,
uranium and potassium4O occurring in the pottery and burial soil, plus a
minor contribution from cosmic rays.
t2l SimD|e build-u
Although an gross oversimplification it is convenient for discussion to use A
this model
t3l Band model
By comparing the 'natural, 'asfound TL with the artificial TL from
portions to which lab doses have been administered one can deduce the dose
of radiation received since last heating. This is called the paleodose and
after evaluating the dose-rate by radioactive analysis of pottery and soil.
the age is obtained: 3Ql Paleodos
The same applies when optical stimulation (OSL) has been used for eviction,
or. when the trapped charge population is measured by electron spin
resonance (ESR).
In the case of TL there is a built-in test of reliability;
t4`| Plateau test
With respect to the plateau one should note that among other things it N
shows that the sample was well zeroed at time zero; also that above about
300"C the traps responsible for the TL have an adequate lifetime for
electron retention. Below that temperature there is TL from the
laboratory-dosed portion but not from the 'as-found' portion; above that
temperature the shapes of the glow-curves are a good match to each other.

Evaluation of dose-rate has many complications, particularly because one is

dealing with a mixed field of radiation in an heterogeneous medium
s5 I oz. li. V
Because the gamma component comes from the soil careful attention needs
to be paid to the surroundings of the sample
t5l_FOCR Output
 ciizn/2 to 3
and because of the short range of alpha particles the dose-rate received by
a grain is dependent on its size
s6l Ouartz drain

Tl. dating was developed for pottery but in archaeology its extension to
other materials has been of greater importance because Tl. can reach back
beyond the 50,000 year limit of radiocarbon. Being able to do this
overshadows its rather poor precision.
t6] Phases of Tl.

Flint has been of dominating importance, being heated in antiquity on

Palaeolithic sites either through accidentally falling into the fire, or,
deliberately in order to improve its cutting properties as knives and
spearheads s7] Photo
.. lt has excellent Tl. characteristics; also it is impervious to moisture (which
otherwise needs to be measured on account of attenuation of doseratei.
0f comparable importance on Palaeolithic sites is the dating by ESR of tooth
enamel:

t7l Tooth enamel for ESR

The outstanding impact made by flint in respect of human origins, has been
made particularly through the work of Helen Valladas at Gifsur-Yvette,
using flints found in association with skeletons in caves such as this one
(Tabun) in Israel $8] Tabun photo
and showing the skeletons there to be of substantially greater antiquity
than had been coniectured.
t8l Tabun dates
The work of Helen Valladas came into worldwide prominence in respect of
whether Modern Man s9] Modern man
developed from Neanderthal l\/lan, which was the established view
sl0] Neanderthal
By dating flints associated with the skeletons of each
t9] Kebara & Qafzeh dates sl l] plus bones
she showed conclusively that the established view was wrong, and that
anatomically Modern Man was in Western Asia at least by 100,000 years
ago, long before the Neanderthals died out around 30,000 years ago. This
gave strong support to the "Out of Africa" hypothesis, developed from
genetic research, that all modern humans are descended from a small
number of people that emigrated from Africa only several hundred thousand
years ago.
Of course the early dates for Modern Man have been questioned. One
possibility investigated was that the external gamma dose-rate during the OCR Output
 cmu/2 to 4
burial period was substantially higher than that measured today. However
this was easily dismissed by means of the isochron technique, well
t10] lsochron established in geological dating
likewise the suggestion that the flints had not been well zeroed because of
insufficient heating --- dismissed because the plateau was good. Agreement
with subsequent ESR and uranium-series dating on this and other sites.

Burnt flints are rare on some sites and the dating of unburnt sediment is
then important, whether by TL or by Opticallystirnulated Luminescence
(OSL). Dating of sediment is also of increasing importance in Quaternary
research. From NW Europe to China there are vast deposits of loess, as well
as in many other parts of the world. This is excellent material for
luminescence dating because of the long exposure to sun/daylight prior to
deposition, giving good setting to zero
t11] Glacier & Lake
An example of a deposit at Biache in N.France, where the 200,000year-old
sediment contained a preNeanderthal skulls $12] Biache
Likewise sand dunes can be dated, here illustrated by the highest dune in
Europe on the Atlantic coast of France, near Bordeaux $13] gg
OSL is advantageous because setting-tozero of the signal is much, much
better. This because the TL signal comes from unbleachable traps as well
as bleachable traps; by definition the OSL signal comes only from the latter.
t12] Bleachable traps
t13] Resettin
The technique was pioneered by Huntleys group at Simon Fraser University,
in Canada, using green emission from an Argon laser for stimulation
t14] Huntley $14] Laser beams
An unexpected development came from Galina Hutt in Estonia who showed
that with feldspar deep enough traps could be stimulated with infrared
$15] Hutt spectrum
The mechanism proposed involved a two step process of eviction: (i) optical
excitation to an excited state about O.1 eV below the conduction band; and
then (ii) phonon excitation from that state into the conduction band.
15] Hutt model
In practical terms this meant (a) that for feldspar dating the laser could be
replaced by an array of a dozen infrared emitting diodes; $16] Diode arra
and (b) a feldspar-only signal could be obtained from a mixture of quartz
and feldspar grains, thereby obviating the tedious business of mineral
separation. OCR Output
 CERN/2 to 5
The application of sediment dating having greatest impact so far has been in
N.Australia, s17] Australia
showing that man's arrival there was around 50 60,000 years ago,
t16] Australian dates
rather than the 40,000 years indicated by radiocarbon (near its limit, where
it is highly susceptible to modern contamination). A number of other sites
give the same indications.

ln conclusion, l will just illustrate an application in which TL immediately

had a very strong and dramatic impact: in testing the authenticity of art
ceramics. Since it is only a matter of distinguishing between ancient and
modern a very clearcut answer is usually obtained, as in this first
example. $18] Weazel & Zebu
The Han weasel had a natural TL comparable with the artificial TL induced
by a lab dose appropriate to its age of a bit over 2000 years; but the
supposedly Persian Zebu had a natural TL minuscule in comparison with an
age of only 100 years.
The samples are obtained by drilling about 100 milligrams of grains from an
unobtrusive part, e.g. underneath the paw. s19] Drilling
This supposedly 2000-yearold Etruscan lady, Diana the s20] Dgga
huntress, residing in the St.Louis Museum, turned out to be a mere 30 years
old, consonant with her evident beauty.
One of the most comprehensive programmes we have undertaken has been in
respect of Hacilar Ware, supposedly made in SW .Turkey some 6 or 7000
years ago. Besides goddess figurines there were, some in major
international museums (British Museum, Metropolitan New York), half-a
dozen magnificent double-headed jugs, with obsidian eyes s21]Fake Dl~l
Certainly magnificent in appearance, but puny in terms of TL
t17] Fake glow-curve suggesting an age of 100 years or less.
Eventually we found one with TL consonant with its supposed antiquity
t1B] Genuine and this one now resides in a museum near Berne
s22] Photo (genuine)
Of course there were claims that the TL had been removed by recent heating
but exhaustive tests showed this not to be the case.
Our results are not always denigratory. Art historians had doubts about the
authenticity of this supposedly Etruscan lady from the Museum of Art and
History here in Geneva s23] Etruscan head
The TL was solid in her support but nevertheless the art historians were
convinced otherwise and suggested that the head had been clandestinely
irradiated with gamma rays. ln this case we were able to dismiss this
possibility by measurement of magnetic remanence, as l will be discussing
tomorrow. OCR Output
 CERN/2 to 6
Restricted Biblioqrabhv & References

Aitken, M.J. (1985) Thermoluminescence Dating. London: Academic Press.

Aitken, M.J. (1994) Optical dating: a non-specialist review. Quaternary
Geochronology 2 3, 503-508.
Aitken M.J. & Valladas H. (1993) Luminescance dating relevant to human
origins. ln: M.J.Aitken, C.B.Stringer & P.A.Mellars (eds.), The origin of
modern humans and the impact of chronometric dating, 27-39,
Princeton University Press, Princeton, New Jersey.
Fleming S. J. (1970) Thermoluminescent dating: refinement of the quartz
inclusion method Archaeometry 12, 133-145.
Fleming, S..l. (1979) Thermoluminescene Techniques in Archaeology.
Oxford: Clarendon.
Grun R., Stringer C.B. & Schwarcz H.P. (1991) ESR dating of teeth from
Garrods Tabun cave collection. Journal of Human Evolution 20, 231
248.
Huntley D.J., Godfrey-Smith D.l. & Thewalt M.L.W. (1985) Optical dating of -OCR Output

sediments. Nature 313, 105-107.

Mercier N. & Valladas H. (1994) Thermoluminescence dates for the
Paleolithic Levant. ln Late Quaternary Chronology and Pa/eociimate
of the Eastern Mediterranean ( Eds.O.BarYosef & R.S.l<ra)
Radiocarbon 7994, 13-20
Mercier N., Valladas H., Valladas G., Jelenek A., Meignen L., Joron JL. and
Reyss .lL. (1994) TL dates of burnt flints from Jelenek's
excavations at Tabun and their implications. Journ. Archaeological
Science, in press.
Roberts, R.G., Rhys Jones & Smith, M.A. (1990) Thermoluminescence dating
of a 50,000-year-old human occupation site in northern Australia.
Nature gg, 153-156.
Roberts R.G., Rhys Jones, Spooner N.A., Head M.J., Murray A.S. & Smith M.A.
(1994). The human colonisation of Australia: optical dates of 53,000
and 60,000 years bracket human arrival at Deaf Adder Gorge, Northern
Territory. Quaternary Ge0chr0nology1 3, 575-584.
Valladas, H, Joron, .l.L., Valladas, G., Arensburg, B., Bar-Yosef, O., Belfer
Cohen, A., Vandermeersch, B. 1987 Thermoluminescence dates for the
Neanderthal burial site at Kebara in israel, Nature 330. 159-160.
Valladas, H, Reyss, .l.L., Joron, .l.L., Valladas, G., BarYosef, O. &
Vandermeersch, B. 1988 Thermoluminescence dating of lvlousterian
Proto-CroMagnon' remains from Israel and the origin of modern man.
Nature 331. 614-616.
Valladas, H. 1992 Thermoluminescence dating of flint. Quaternary Science
Reviews 11, 1-5.
Zimmerman D. W. (1971) Thermoluminescent datin`g using fine grains
from pottery Archaeometry 1 3, 29-52.
 CERN/2 LD 7
CERN Lecture 2 Mar21,'95 LUMINESCENCE D/\T1NG:
(t = transparency) (S = Slide)

/t1] Simble dlow-curve

$1] Nok head
$2] Apparatus
$3] Bovle

t2] Simble build-u

t3] Band model
$4] Paleodose.
/ t4] Plateau test

S5] oi. B. Y
t5] Pit

$6] Ouartz drain

t6] Phases of TL

$7] Photo
t7] Tooth enamel for ESR
$8] Tabun photo
t8] Tabun dates
$9] Modern man
5.iO] Neanderthai
t9] Kebara & Qafzeh dates $11] plus bones
1/ t10] lsochron

t11] Glacier & Lake

$12] Biache

t12] Bleachable traps

t13] Resettin
t14] Huntle $14] Laser beams
S15] Hutt spectrum

t1 5] l-lutt model glu bmah]

t16]Au$tralian dates $17] Australia

$18] Weazel 8 Zebu

$19] Drillin

t17] Fake gl0wcurve $21]Fake D1l

t18]Genuine $22] Photo (genuine OCR Output
7} $23] Etruscan head
/ .

Q/IM
 bssvd sjznpg)
.-7)/VJDS`,7/V/Nfl ww A/3H.L OCR Output
 /

\ Present ground surface

x BAD
`
Earlier ground surface
/ Ancient around surface

\\ J Pooe
L { /

`
/\

$1 POOR
\/
\_/

1
L {_
- i OCR Output

me 1..1 OCR OutputOCR OutputOCR OutputOCR Output2) A4!/K

 79//Ass
g . 72. LAT//*6
PHASEI ARCHAEOLOGICAL POTTERY
AND ART CERAMICS
(TL REVOLUTION)

PHASEII Beyond Carb0n14 <40.000 years>

BURNT FLINT
VOLCANIC LAVA
<fe|dspar peak at ~6OO C>
STALAGMITIC CALCITE
<in paleolithic caves>

PHASEIII;UNBURNT SEDIMENT

<reset by sun )(> \

WINDBLOWN WATERBORNE
OSL
=x.e1=RR@| 1.0Ess DUST 'N
GLACIERS

SAND GLACIAL SEA

DUNES LAKES FLOOR

QUATERNARY GEOLOGY OCR Output

 OCR OutputTABUN CAVE, ISRAEL: TL d BL/RNED FL//V7'
V,LLcca.: eCaJ,. /9;%
I__CTr_H_
3 BI 5 9 % 11 % Tw$c.AL

IX

XI

XII

XIII

100 200 300 l00

{ Age(k) THOlS;g&<OCR Output

4\.A.n.
 NEA NBER THA; Mobek/v M
bg Ungt X// nv ALL LAYERS

KEBARA QAFZEH
Init Layer
VI

XVII

VII

XVIII

VIII

XIX

IX

XXI

XXII
XI
XXIII

XII

50 IOO

Age (thousands of years)

F6. I3 Burnt Hint dates tbr two caves in Israel. A Neanderthal skeleton
Fzbntly found in Unit Xll ot`Kebara Cave and `Proto-CroMagnon remain
een found nt Qatkeh throughout the layers shown. (Deriv OCR Output erI *`- "" '
 GKEEN uqyf
am Lasse
OCR OutputOCR OutputOCR OutputBLu
LuMuvsc~c
/ / / / / <
"""'*""" //

BLEACJIABLE
uuuen annex.;
BOTH TRAPS
Lumwesemez crvs 7;; szc;vAz..
canvas

Fok .DA T wg- : .D7'cv7o/v wnvasuqfrx

ST/MULRTION
wnvL=1.;vc;vw
Le. HNT/S7OKES

,5/MPL/F/5) NOLEL
OPT/CAL LY - ST/M um TE) Lu/wxvsc/vc
{cse)
$ OCR Output070-$UMuLATEp Lz/M. (7?5.L_ HPL.
 I

<- CEOLOC/CAL.
/GN/QL
BL EACH//\/Ci
U/V

TL

OPT /041.
SW/*10;. =m0A/
05 .4.

[ARS uk 1AL

U/V EXPOSURE /\/EE`bEb .-

~ /0 A/00/QS

/A/uT$ O S L OCR Output

INFRARED STIMULATION 0f
KFE|.DSFAR

(||C1tt 8. Jcnek 1989)

930 nm

870nm

550nm

`>=X..,.,.,<.r<f
1-5 2O 25
Stimulation energy (eV) OCR Output
 OCR OutputOCR OutputOCR Output.-b
...IO

O O
C @ LD y Q G
A.LISN31.N| 'I.L OCR Output
 OCR OutputOCR OutputOCR OutputOCR OutputOCR OutputOCR OutputOCR Outputy J<#>wMJ "~4~*-:7 A/'*f.~/'V1""V*- '0.j/ y/1_Sy*-+*7r"* "vrw-uuj ~/ 1rt,.fiv.>1
P`c /
TL
CERN/2 tu 7
m
_,, t1 Sim le low-curve bj * " s11 Nok head

K / $7- ne g"j" fwnp/(K 5444 wm? $2] Apparatus!] [Q

i""$ ""$ff"{i ml? _,,,\ 0[ae1*-5m0+ , EOOWE/"iZi<LYl
EEE
t2| Simple build-u
gp t3] Band model @p,yp,y,? QQ)
OSL ~
_ W [SR - -7'>07!f ENAMEL- A7;,;/Cgif; /
Paleodose. >/in
_/ t4| Plateau test ---.3
4 5 A
SMTUA. ..... W
S5]cx. 15. v
t5] Pit _
$61 Ouartz drain

Y6 ] BE}/ON`/5 EAQIO BO/Y VAL!-f=j>AS

V1/ft//1A7V OKI;/A/S
%__{Qj{__ F1. lr

`[p_dtl]_enamel for SQCTL QW TL $7] Phctc

QOH/7 t8l Tapun dalgs wm? @8] Tabun pn0t0
_ _ A
" it $K
{_ _ _ _ .2 A _ _ 1,; SQTI]/lOdlTl man
A
s10] Neanderthal

./ UO]
t9] Kebara & Qafzeh dates 'j;';'F"%*;JL;L%.. $11] plus bones iO.*;*USl _

Exlirtml +6% ESR QL U,S/_H;$ qRAciz.

t11] Glacier & Lake iB~lLf>lA/ytL5m~t

Seclumemlf 9 s12] Biache
_ A _ _ __ __
S13] Byte
n "`
tlg] Bleachable tr 0p;,c;,_ ~ \ _ _,_
? U2,] Resetting
I ] ` t1'] Huntley ek $1] Laser beams
Cam. IR $15] Hutt spectrum
.. ,_ t1 5] Hutt model C ~..~,-c f" d`P
` _ 2f> i/MEA? S Igl blalb
l/`(tli] Australian dates $1lZq Au$tralE

$18] Weazel & Zebu

$19] Drillin

t17] Fake glow-curve

t18]Genuine *__ $22] Phote (genuine
RNE YMMBE , Q $23] Etruscan head
nolb 1>1/@94 , wi/ik MAKE) MEN OCR Output

al uJ:j$ 7 . CLMV ES" VE D { {I X

DEKOQ AJORY
 M ZS LUM/NESCENCE .DA'n/V

ER/WOLUM/NESCENCE /71/,

-Qt/ /0/7/'Vln 0c ue

u:l 0h.L

Home sapiens sapdens

QI/6/Q In /Vd.7ld@i*

0+ M1,

bi/LLJZQ Q] w71DufMz 5BdL/77-Q'!

and

S//7hu..Lde,
LLfnLl1SCht
s0.' Sbelm wa well anim

MVa,L cj mam bh Hoes

OLKZ err q@YLbbTl, I

SLU'!. diff GZ7/h1bC.S OCR Output

 CERN/3: Maqnetism & Arcnaeoloqv

CERN Lecture 3: Mar Z2,'95 : Magnetism in Archaeolo<g A/OES

(t = transparency) (s = slide)

There are several aspects to the involvement of magnetism in archaeology:

tl] ltiagnetism in Archaeology:

(i) Pottery kilns, hearths etc.: Direction of earth's magnetic field during
last few thousand years, possibility of use for dating
(ii) Pottery, bricks and tiles: Intensity of earths field over last 8,000 years
(iii) Sediment: intensity over hundreds of thousands of years: use for
carbonli calibration

(ix ) Sediment & lava: Reversal stratigraphy. millions of years

(v) Location of buried pottery kilns. pits and ditches
(vi) Climatic indicator: magnetic susceptibility changes in sediment

First of all, let me remind you: the earth's magnetic f ield. as observed on the
surface of the earth can be quite well described in terms of a bar magnet at
the centre sl] Earth as bar magnet
The elements of the field being:
Strength. or intensity, F
Declination (deviation of magnetic north from geographic north), D
Inclination, or Angle of Dip, I
Of course, the bar magnet is only hypothetical: s2] Earths core
the major part of the field is generated by electric currents in the fluid part
of the core, the currents being driven by convection.
From observatory records there was indication of a 4 or $0O-y ear
periodicity, as though the hypothetical bar magnet was inclined to the
earth's axis of rotation and precessing: s3] Bauer curve

ln baked clay there is a record of the ancient field at the time that the clay
cooled down from baking and so dated pottery kilns etc enable us to see
further back in time. The magnetic record is carried byrhermoremanent
magnetization (TRM) and is due to a small degree of aiigneinent of the
magnetic domains in iron oxide
si] TRM of baked clay
Alignment takes place while the clay is hot; on cooling it is frozen and this
fossilized magnetic moment, athough weak, is remarkably stable --- over
hundreds of thousands of years.
The result is that a pottery kiln carries a magnetic record
si?] Kiln with arrows OCR Output
 Q s?] Kiln structure
it is the kiln strucure that is sampled for direction, but the pots are useful
lor intensity. 311, stl] Frames
46 $$3*] Piasterin
The frames are levelled and the direction of geographic north marked on the
top surface of the plaster (using a theodolite).
T shall come back to results obtained shortly, because you must first find
your kiln and this is where magnetic location comes in.
There is about a ton of clay in a pottery kiln and so, although the TRM is
weak the total magnetic moment is substantial and there is a small but
significant distortion in the earths field at the surface slO] Cartoon
The first technique used for detection was based on proton free precession
the frequency of precession being proportional to F.
A sizeable sample of protons is conveniently available in a bottle of water
sl ll Bottle
and in the early pre-automation days one operator would move the bottle.
encapsulated in resin, while the other recorded the reading indicated by the
electronics sl2] Survey in progress
ln this way a magnetic chart was obtained, at the rate of about halfa
hectare per sl3] Stibbington chart
in addition to the disturbance due to a kiln there is also a linear one due to a
ditch. This is the kiln we found sit] Kiln photo
All that was back around 1960 and in the first big survey we did we
discovered what is now taken for granted -- that a filled-in pit can give as
big a disturbance as a kiln sl5] Pit photo
This is because the magnetic susceptibility of the pit filling is high than that
of the subsoil or rock into which it is dug. This enhancement is due to
conversion of iron oxide from haematite into maghemite. For the same
reason a buried wall can be detected slo] Wall photo
though the disturbance is weaker.

Returning now to the results obtained from measurement of TRM direction.

let me remind you that when this work was started by Thellier in the l940s
there was still expectation of of cyclic behaviour, as though the hypothetical
bar magnet; was precessing with a periodicity of around t or SOO years
But the archaeomagnetic results showed distinctly otherwise, as illustrated
bv the curve we obtained for Britain sl?] Q;
The dating is from archaeological chronology and this is a prime example of
the two-way flow of benefit between archaeology and physics. or rather OCR Output
 CERN/3: Maqnetism & Arcnaeoto
geophysics. But also, a limited amount of dating of pottery kilns of unknown
age can be done.
In those early days the measuring apparatus was rather larger than now
and this enabled us to measure the magnetic direction recorded in terra
cotta statues such as this Chinese garden gnome sl 8] Gnome photo

More recently it has been possible to turn attention to ancient intensity. F.

and utilize quite nice looking Greek pots as a source of sample
sit?] Pot photo
To find ancient E you have to heat up to 700"C (and then remagnetire in lab
field); obviously museum curators would be unhappy about this. However
by utilizing a SQUID cryogenic magnetometer
s2t>] SOLIID
it is possible to work with a very small sample s2l] Core on stalk
f
The most interesting thing that hgglfchiie out of this work is that between
ZOOO andlOOO BC there was a strong rise in intensity, reaching more than
50% higher than present day. s22] Early Near East
s23] Greece
At present the error bars are too large to think of this as a feasible dating
method; however, like the directional results, it is of considerable interest to
geophysicists seeking to understand the mechanism by which the earths
field is generated. Results from other parts of the world, from China round
to N.America, indicate that the strong peak was not synchronous but drifted
around slowly from East to West.
I should mention here that work is in progress at Liverpool to use
microwaves for the remagnetization process and this holds high promise to
improve the precision of the technique. \-\-ith microwave heating you can
reset the magnetization without heating the sample: this ay r oids the
mineralogical changes that interfere when ordinary heating is used.
s24] Etruscan Head
We now return briefly to yesterday's Etruscan lady whose integrity had
been impugned by art historians despite TL having vouchsafed her a clean
bill of health. You may remember that it was alleged that she had been
given a false TL age by clandestine irradiation. mention being made of the
italian mafia as agents in this. Well, knowing that the magnetic intensity in
Etruscan times was substantially higher than today we submitted a small
fragment of her to intensity determination s2 5] E values
She did indeed carry indication of the high Etruscan intensity, some 7-t
microtesla, substantially higher than the 41 microtesla carried by known
fakes. Of course the suggestion was made that she was a forgery OCR Output
 CERN/3: Maqnetism & Archaeolo
nevertheless, having been baked in a part of the world where the intensity
is high today. Well looking at a present day chart of worldwide intensity
s26] Worldwide F
we see that you find 74 microtesla in Siberia and Northern Canada, hardly
likely palces for the Italian mafia! On the other hand. in respect oi` the 41
microtesla indicated by the known fakes it is interesting to ask where in the
world the present day field has this value oddly enough one of the
regions is centred on Sicily.

in conclusion l will mention the Geomagnetic Rex i~ ersal Timescale. ifolcanit

lava acquires a TRM in the same way as baked clay; rather supprisingly
sediment deposited from water (e.g., on the ocean floor, on lake beds) cities
so too. This Detrital Remanent lvlagnetization is very much weaker but
nevertheless is stable. It is acquired due to alignment of the magnetic grains
as they settle or during the consolidation process after deposition. From long
cores taken from lake bottoms a continuous record of D & i can be obtained.
and from that taken from lakes in Britain (Windermere. Lomond etc.) we get
good confirmation of the data obtained from kilns. But the more important
aspect of sediment magnetization is being able to relate early hominid
remains to the Reversal Timescale. Roughly every million years the eartlrs
magnetic field flipped over; these are the major chrons and within each
there are minor events in which the reversal only lasted for around 10.000
years or in which it failed to establish itself. The dating of these is based
mainly on potassium-argon determinations on lava in which reversals are
recorded; some of the major chrons and subchrons of the last 2 million years;
t2] Reversal Timescale
For sites on which there is a long enough sequence of sediment for this
timescale to be displayed then the dating of any remains is clearc ut. W here
only the sediment surrounding the fossil is available then the magnetic
polarity recorded in that sediment acts as an important check on the age
obtained using a radiometric technique. An example of this is the i{.>\r ages
for the skullcap of Homo erectus child, and another specimen, in java
published last year.
t3] [ava H.Erectus
Accepted ideas on H.Erectus were that he ventured forth from Africa about 1
million years ago: the KAr dating of hornblende crystals associated with the
skullcap gave 1.811 0.04 million years. However this is concordant with the
normal polarity found there, placing it in the Olduvai subchron within the
reversed lviatuvama Chron. OCR Output
 CERN/3; Magnetism & Archaeolo

few references

Aitken M.J. (1990) Science-based dating in archaeology . Longman, London

and New York, 274 pp.
Bucur l. (1994) The direction of the terrestrial magnetic field in France,
during the last 21 centuries. Physics of the Earth and Planetary
interiors 87, 95-109.
Cande S.C. & Kent D.V. (1992) A new geomagnetic polarity timescale for the
late Cretaceous and Cenozoic. Journ. Geophys. Res. 897, 13917~13951.
Swisher C.C., Curtis G.H., Jacob T., Getty A.G., Suprijo A. & Widiasmoro
(1994) Age of the earliest known hominids in Java, indonesia, Science
263, 1118-1121. OCR Output
 @@1/;_I\/Iagnetism & Archaeolo
CERN Lecture 3: Mar 22,*95 : Magnetts111j.;ggrgiiaeo1t>gx:
(t = transparency) (5 = slide)
t1] Magnetism in Archaeologvz

s1] Earth as bar magnet

52] Eartl1's core
53] Bauer curve
54] TRM of baked clax
55] Kiln with arrows
56] Pots in kiln
57] Kiln structure
58] Frames
59] Plastering
510] Cartoon
511] Bottle
512] Survev in progress
51 3] Stibbing ton chart
51-1] Kiln photo
515] Pit pl1oto
516] Wall photo

$171
518] Gnome photo

519] Pot photo

521] Core on stalk

522] Earlv Near East
523] Greece
`
524] Etruscan Head
_ @ L/{. 525] E values
I", I
526] \\.'o1ltix-vieie F

t2] Reversal Tiiuescale

t3] iava H.Ereetus OCR Output
 A S CHAMP
BLA Ks
/ V A '

NORMAL wuz BR UNHES

REVERSE} R
AQE5 n
0 gig
M/LL/ONS
0-98 4 ya/ms
TRRA M/Ll. O -+
05
COBB

NATUYANA
/79
OLDUVAI >
/~95

Reuxwoxv .> 2/1+

2-6
GA USS OCR Output
 TA VA
_ jgg
hlf
//O/*70
Rcvus OCR Output

%D
 /
P/0/70 EREGTUS -{

&,1~(&cs' {dm'. uh AFR/CA. $3 Md

//=
,, ) Jn ven bafr szm, /60/S
/-4- Ma
C,u,LU.cve

C/ui
Ladd

7(6 x M0; /.2% E"`/

fs JL" wa.: CL /0n,q {ima Z>e.:v*e
Aa S AFr2
Bu,./Y who we rz0 Acmlmn falg

A/OW %/0 m/ende

of ' /;*.5. _ 0%. Java
/`?~m V1 xvoemaz. /2/z 0-0
MQ
K HE Se./T'0 $0m H'-Pnl
z* /0.n

veh cme da /iv-.v~ E8? OCR Output

A
 OCR OutputOCR OutputOCR OutputLUIHIHCSCCHCC d3t1I1g I`ClCV3.I'1t tO hl1IT13.1'1 OI`1g1I'1S
M. j. AITKEN1 AND H. VALLADAS2
'Reseansh Laboratory for Archaeology, University rf Orgford, 6 Keble Road, Orjord OX] BQ.], U.K.
, 2 Centre des Faibles Radioactivits, Laboratoire mixte CNRS-CEA, Avenue de la Terrasxe,
91198 Gy' sur Yvette Cedex, France

SUMMARY

Luminesccncc dating provided thc first direct and indcpcndcnt cvidcncc that anatomically modern
humans had a prcscncc in wcstcm Asia carlicr than is consistent with thc rcgional c0ntinuity model.
The reliability ofthe result concerned, 92 (j 5) ka for burnt Hints from Qafzch Cave, is excellent and
consistent with isochron analysis ofthe data. Flint dating has also confirmed palaeoenvironmental
indications that the Mousterian industry in Europe was present somewhat earlier than the 100 ka limit
previously accepted. Burnt quartz and unburnt sediment have also been important in Palaeolithic
dating and the latter has a particularly high potential.

1. INTRODUCTION
(41) Basis
The two commonly used techniques of luminescence
dating are thermoluminescence (T1.) and optically
stimulated luminescence (ost.), the latter being called
optical dating by its originators (Huntley et al. 1985).
For T1. the dating signal is stimulated by heat whereas
for ost. it is stimulated by light. X/Vith both, the signal
is a measure of the population of electrons trapped at
defects in the crystal lattice of the mineral being
utilized (e.g. quartz, flint, feldspars); the buildup of
this population is the result of continued exposure to
the weak flux of nuclear radiation emitted by radio
active impurities in the sample and in the immediately
surrounding sediment, together with a minor contri
bution from cosmic rays. The relevant radioactive
impurities are 232Th, 2351.1 and BSU together with their
associated radioactive decay products, plus 4UK and
7Rb, although the latter is of almost negligible
importance. The relevant nuclear radiation consists
efectively ofot and B particles from within the sample,
y radiation from the burial surroundings {up to a
distance of about 0.3 m) and cosmic ray mesons.
For the trapped electron population to be a useful
measure of age it is essential that it was zero at some
event in antiquity; it is that event which is dated. This
zeroing or resetting can be by heating to upwards of
around 400C or by sufficient exposure to daylight. If
pottery is under study then the event will have been its
Bring in the potters kiln. In the case of burnt flint it
will have been accidental falling into the fire or
deliberate heat treatment. For aeolian deposits such as
wind-blown sand and loess the resetting event will
have been exposure to light during transportation or
while lying exposed on the surface before being
covered by further deposition. Waterborne deposits
will also have been reset, though not so effectively; the
ost. technique is particularly advantageous for such
sediment because with this technique the dating signal
is stimulated only from highly light-sensitive traps
whereas the T1. signal comes from other traps too,
some of which require hours of short wavelength
bleaching for their effective emptying; also there are
some traps which are virtually immune to light
exposure. This is in contrast to the brute force
zeroing that results from heating. Trapped electron
populations of zero are also the case for newly-formed
crystals e.g. stalagmitic calcite.
To translate the dating signal into calendar years
two further quantities are required. First, for each
sample studied it is necessary to measure the sensiti
vity, i.e. the signal resulting from a given dose of
nuclear radiation; this is done by exposure to artificial
radiation of known intensity. In this way the palaeo
dose, P, can be evaluated; this is the laboratory
estimate of the dose that the sample must have
received during antiquity for its dating signal to be
equal to the observed value. Secondly, it is necessary
to determine the dose-rate, D; this is the dose per year
that the sample has been receiving during its period of
burial. One approach is to determine the concentra
tions of radioactive elements present in sample and
soil by means of neutron activation analysis. Alterna
tively, direct measurement of the natural radioactivity
is used, both within the laboratory and on-site. The
cosmic ray contribution is obtained by calculation,
having regard to thickness of overburden, altitude and
latitude. In principle the age, A, is then obtained from
the equation.
A = P/D.
The unit of dose is the gray (Gy). The age so obtained
is directly in calendar years and it is independent of
any other chronological technique.

Phil. Trans. R. Soc. Lond. B (1992) 337, 139-144 139 @ 1992 The Royal Society and the authors
, _ Printed in Great Britain OCR Output
 In practice, cquation (1) is deceptively simple and
even with automation the derivation of a reliable
result is complex and labour intensive; a dozen or so
parameters need to be measured to make allowance
for various subtle eH"ects: accounts of these have been
given elsewhere (Aitken 1985, 1990). It should be
noted that in contrast to the radiocarbon and ura
nium-series techniques the radioisotopes involved
have very long half-lifesin excess of10 years; hence
the dose-rate is basically constant (there may be small
variations due to varying environmental conditions).
A measure ofthe trapped electron population can
also be obtained by electron spin resonance (Esta) as
discussed elsewhere in this volume. The three tech
niques are often collectively described as trapped
electron dating. Recent reviews of luminescence tech
niques and applications have been given by Berger
(1988), \Vintle (1990), Zoller & Wagner (1990),
Aitken (1989, 1992) and Yalladas (1992); the solid
state mechanisms involved have been discussed,
among others, by NlcKeever (1985).
2. RELEVANCE TO HUMAN ORIGINS
(a.) Burnt fint; burnt quartz
Flint is a form of chalcedony, another being chert; in
the context of dating the two are not usually dis
tinguished, both being called flint. Its dating came
into prominence with the publication by Valladas et
al. (1987, 1988) of results {or two caves in Israel:
Kebara and Qafzeh. The result for the latter gave the
first hard evidence that anatomically modern humans
were present in that region some 90 ka ago, thus
firmly rebutting any notion of evolution from Nean
derthals; the cave at Kebara had contained a Nean
derthal skeleton and the TL dates for the relevant
levels indicated an age of close to 60 ka. These were
subsequently supported by ESR. Recently another T1.
age indicating early arrival of modern humans in that
region has been obtained at the cave of es-Skhul
(Mercier et al. 1992). The reliability ofthe Qafzeh
date is examined in 3d.
Another contribution, recently reviewed by Valla
das (1992), has been confirmation of palaeoenviron
mental indications that the Mousterian industry in
Europe was present somewhat earlier than the 100 ka
limit generally accepted. Thus at Biache-Saint-Vaast
in northern France, TL dating of six burnt flints gave
an age of 175 ( j 13) ka for the level in which two pre
Neanderthal skulls had been found: floral and faunal
evidence indicated occupation during an interstadial
of isotope Stage 6 (Huxtable & Aitken 1988:1).
Another site for which rt. gave an earlier than
expected age for a Mousterian (IV) level was at Abri
Vaufrey in the Dordogne valley of France; four Hints
gave an average age of 120 (j 13) ka (Huxtable &
Aitken 1988b); a similar age had previously been
obtained for a stalactite fragment found in the same
level (Aitken & Bussell 1982). The oldest site for
which flint dates have so far been published is the
Acheulian site of Maastricht-Belvedere in eastern
Holland, the age obtained for the earliest occupation
layer being 263 (j 22) ka (Huxtable & Aitken 1985).
Phil. Trans. R. Soc. Lond. B (1992)
Another silica mineral utilized is quartz and grains
extracted from the clay of Palaeolithic fireplaces have
been used for dating: ages of 31-36 ka have been
obtained for the Lake Mungo fireplaces in Australia
(Huxtable & Aitken 1977; Bell 1991). These are in
satisfactory agreement with radiocarbon if allowance
is made for that techniques probable underestimation
by several thousand years (Mazaud et al. 1991).
(b) Unburnt sediment
Although burnt Hint is an excellent material for TL
dating, a severe problem on many sites is the scarcity
of Hints which are both well enough burnt and large
enough for satisfactory processing. Hence the feasibi
lity of dating the time of deposition of unburnt
sediment vastly extends the scope of luminescence
dating on Palaeolithic sites. An early study of sedi
ments from Garrods Tabun Cave was made bv
Bowman (1985). Although this was on polymineral
samples there was indication that TL from quartz was
dominating the signal, the problem of feldspar fading
hence being alleviated. Bowman (1985) was reluctant
to quote a date on account of doserate uncertainties
associated with observed escape of radon from samples
in the laboratory. Nevertheless it is interesting to note
that using the data reported, the ages obtained
(making approximation about radon loss; S. G. E.
Bowman, personal communication) for two samples
from Garrods level D (the level below the female
Neanderthal skeleton) were both about 160 ka, close
to the linear uptake (LU) age of 166 (j 20) obtained
for tooth enamel from that level by ask (Griin et al.
1991). On the other hand TL dating (Mercier 1992) of
11 burnt flints from level D has given a preliminary
average age of 250 ka. Of course in dating sediments
in dark caves it is essential that the grains were carried
in by wind (or water) rather derived from the walls.
Another relevant study is that of Roberts et al.
(1990), made as the first in a series concerning the
time of initial human arrival on the Australian
continent. Quartz grains extracted from sand on sites
located on sand aprons at the foot of the Arnhem
Land plateau were used and some dozen dates
obtained at various depths in the profiles. There
was satisfactory agreement with radiocarbon ages
obtained for the upper layers of both sites, and at one
of them three TL ages were associated with the lowest
occupation level yielding stone artifacts; the authors
suggest that human arrival in northern Australia
occurred between 50 and 60 ka ago, somewhat earlier
than previous indications obtained using the radiocar
bon technique (which is of course at its limit around
40 ka).
In using TL for unburnt sediment the need to assess
the effectiveness of resetting at deposition adds sub
stantially to the work. The OSL technique avoids this
and highly encouraging results have been obtained in
test programmes (e.g. Smith et al. 1990; Stokes 1992).
An age of around 130 ka has been obtained (Stokes
1993) for quartz grains extracted from sediment of the
lowest lake at Bit Tafawi (Egypt) in agreement with
uranium-series and amino acid dating of ostrich
eggshells. OCR Output

(c) Stalagmitic calcite
Although T1. can bc used for dating this type 0f
material in caves occupied by hominids back to 300 ka
(eg. at Caune de lArago, France, by Debenham &
Aitken (1984)), in most contexts the heterogeneity of
the calcite itself and of the surroundings means that
reliable assessment of the external dose-rate is difficult.
The uranium-series method does not have this draw
back and hence is usually the preferred technique.
The calcite date mentioned above for Abri Vaufrey
was obtained for a small piece of stalactite which had
fallen from the cave roof and become buried in
sediment to a suiiicient depth (0.3 ml for the external
dose-rate to be reliably evaluated.
3. RELIABILITY
(:1) General
In addition to the concordances with other techniques
noted above, accumulated evidence from a wide
variety of samples and sites has established the general
validity of the technique {see Ancient Tr. date lists,
Bailif, n.d.). Nevertheless, with a particular sample
type or a particular site there may be special sources of
inaccuracy. Hence, however thorough the demonstra
tion that a particular mineral, say quartz, gives
accurate results on a go0d site, it does not prove it
will do so on a site where interfering factors, e.g.
radioactive disequilibrium, are present; conversely
failure in one particular site or with one particular
mineral does not prove the whole technique to be
invalid.
The dates mentioned in 2 are mainly based on
flint, which has the advantage of being a geochemi
cally stable mineral. In addition to the results
obtained by one of us (HN'.), there has been extensive
flint dating byj. Huxtable at Oxford (see Ancient TL
date lists); all these indicate burnt Hint to be an
excellent dating material. Additional comparisons
with other techniques have been summarized by
Valladas (1992). The oldest was for La Vigne Brun
(France), yielding a Hint date of 27 ka; this was in
excellent agreement with the age obtained using burnt
quartz and also with the average radiocarbon date for
four charcoal fragments after making allowance for
the probable 3 ka underestimation by the latter
technique at the time period concerned (Bard el al.
1990).
Some additional general aspects are considered in
the next two subsections and then there is specific
discussion of the critical date for Qafzeh.
(b) The plateau test; residual signals
An advantage with the luminescence techniques is
that the dating signal is more than a quantity, it has a
characteristic shape too, and irregularities in this are
indicative that all is not well. In the case of TL dating
the signal is in the form of a glow-curve and
comparison of the glow-curve shape observed for
portions of an ancient sample as found (the natural
Phil. Trans. R. Sor. Land. B (1992)
Luminescence dating Aitken and H. Valladas 141
1-0
PLATEAU ',,g\f
/ RATIO
1 os)- {I {A j
{ `~ [ , \ I \ 5 \{SSA
\ ' ` xl
. I `-Lx / M
100 300 500
TEMP! c
Figure l. The plateau test. Curve N is the TL glow-curve for
a Hint that was well-burnt in antiquity and curve A is the
additional TL observed from another portion to which a
laboratory dose has been administered. The ratio (N/A)
reaches a plateau in the glow-curve temperature region
where the storage lifetime of the trapped electrons respon
sible for the TL is much greater than the age of the sample;
at lower temperatures the TL is associated with electrons
ejected from traps for which the storage lifetime is much
shorter. It is the plateau value that is used for dating and in
this example the ratio value (of 0.95) corresponds to the
palaeodose being 0.95 times the laboratory dose adminis
tered. In practice a range of laboratory doses is used, with
several portions for each dose point.
TL) with that for portions that have received labora
tory irradiation allow the all-important plateau test to
be made [see figure 1). Failure to pass this test can be
for a variety of reasons: e.g. contamination, presence
of spurious TL (a parasitic signal indicative of poor
measurement conditions), inadequate stability of the
trapped electrons (i.e. their lifetime in the traps was
not long enough to avoid leakage during the burial
period). Failure can also be caused by insufficient
heating at the resetting event in antiquity and given
the difficulty mentioned earlier of finding enough
well-burnt samples this is of particular relevance to
flint dating (see figure 2). Because the plateau test is
an intrinsic part of the measurement process there is
no risk ofdates being obtained from halfbaked Hints.
This eliminates the possibility of a Hint date being
erroneously too old on account of the dating signal
containing a contribution of residual geological TL.
However, the question of residual signal is of more
serious concern in the TL dating of unburnt sediment;
as mentioned earlier, in such application the residual
signal may be appreciable and needs to be subtracted.
Here again the plateau test is employed, although
usually substantial complication is involved (see Mej
dahl 1988). There is great region-to-region variability
in this respect dependent on mode of deposition and
intensity of daylight, at the Arnhem Land site in
Australia mentioned in 2b the straightforward OCR Output
142 M. Aitken and H. Valladas Luminescence dazing
RATIO
. EA
I\\
1}.. { x
I { `~
A { \\ /// x`
I
I
\\
`~-,L/
O \
\x
100 300 500
TEMP/C
Figure 2. Failure to pass the plateau test. Curve N is thc
glow-curvc from a flint that had not been burnt in
antiquity; curve A represent the TL due t0 a laboratory dose
administered after heating a sample ofthe flint t0 500C.
Curve N has a different shape from that of figure 1 because
in this unburnt case the trapped electrons have been
accumulating during the very long time that has elapsed
since the Hints formation. The data used are taken from
Robins et al (1982).
For partially burnt Hints the rise in the ratio is less steep
but as is evident from the work of Melcher & Zimmerman
(1977) there is no risk of even the semblance of a plateau
with such Hints, making calculation of a date impossible.
plateau was good, consistent with theobservation that
direct measurement of recently deposited sand at the
site indicated a palaeodose of only one gray.
In the case of ost. the corresponding necessary
condition for reliability is a satisfactory shine-pla
teau: i.e. there should be no change in the paleodose
calculated for successive intervals of exposure to the
stimulating light.
(c) Limitations on age range: lifetime; saturation;
anomalous fading
Stable retention of trapped electrons over the burial
period is obviously a basic requirement in respect of
the traps providing the dating signal, and as men
tioned above a satisfactory plateau indicates that this
condition has been met. However, it is useful to have
some estimate of trapped electron lifetimes since this is
a possible limitation in the age range of application.
Laboratory kinetic measurements indicate the fol
lowing rough values (see Aitken 1985): for quartz and
burnt Hint, 108 years; for calcite, 106 years; for K
feldspar, 107 years. Thus we see that for the 100 ka
time span, which is this v0lumes focus of interest, the
lifetimes are adequate.
However, there is another limitation on the age
range. This is that with continued exposure to radia
tion all available traps become filled. For quartz and
Hint the radiation dose at which this occurs is of the
order of several hundred grays, and depending on the
Phil. Tran;. R. Soc. Lond. B (1992)
level of radioactivity this corresponds to a limiting age
of several hundred ka. Some types of these minerals
saturate at lower doses and some at higher doses, so no
strict limits can be laid down; it is a matter for
investigation on each particular site. In general it
seems that Hint can reach further back than quartz.
For feldspars the saturation dose is much higher, a
value of several thousand grays being applicable to K
feldspars for instance. However despite the adequate
lifetime indicated by laboratory studies there is a
tendency in some types for appreciable leakage of
trapped electrons to occur. This is termed anomalous
fading and whenever feldspars are used for dating
special procedures and checks are necessary to exclude
the possibility ofinterference by this malign phenome
non; this applies equally when a mixture of minerals is
used: the luminescence from feldspars is usually
dominant.
(d ) Qafzeh
The date (Valladas et al. 1988) for this cave is based
on 20 Hints obtained from Mousterian layers within a
2.5 m section that had yielded anatomically modern
humans (ProtoCro-Magnons). According to the
multiregional model these humans evolved from
Neanderthals and the date (92 ka) should have been
more recent than the 60 ka date for Kebara. Hence
the question at issue with the Qafzeh date is whether
there was some factor that caused it to be erroneously
(0) 120
100
80
160
(b)
120 >
$4
g 80 11% [if -1
LO
0 0L 08 1-2 1-G
INTERNAL DOSE RATE/Gy k`
Figure 3. (cz) (right-hand scale) TL ages for Qafzeh plotted
against internal dose-rate values. The horizontal line repre
sents the weighted average (as published by Valladas el al.
{1988)). (b) (left-hand scale) Isochron plot of Qafzeh
palaeodose values versus internal dose-rate values. The line
is the least-squares fit, weighted for uncorrelated errors on
both axes (York 1969); the slope indicates an age of
88 (;9) ka and the intercept on the horizontal axis indi
cates an external dose-rate of 0.28 (3; 0.09) Gy ka" OCR Output
 Luminexcence daling M. Aitken and H. Valladas 143
t00 old. The most easily dismissed is thc possibility
that thc Hints wcrc insu$cently heated at time zero;
besides this being ruled out by the considerations of
3b, it is also ruled out by the tight clustering ofthe
individual dates (see figure Ba): the degree of
insumciency would vary from flint to flint.
Given that the paleodose evaluation followed well
established practice other possible causes of overesti
mation have to be sought: could the dose-rate have
I been higher during the millennia of burial than
indicted by measurements made today? The imper
vious nature of flint makes it highly unlikely that there
would have been any leaching away of internal
radioisotopes (which contributed the dominant part of
the total dose-rate) and even if it had occurred it
would hardly have been the same from flint to Hint.
Direct evidence against leaching is that in the Hve
Hints for which fission-track mapping of uranium was
carried out there was practically uniform distribution.
Turning to external dose-rate, albeit of lesser
importance for this site according to the measure
ments reported, there are possibilities that cannot be
dismissed a priori. One is that there might have been
progressive leaching away of radioisotopes in the
burial sediment. Another is that the moisture content
of the sediment might have been lower during anti
quity than at present: the presence of moisture causes
attenuation of the Hux of ,r radiation reaching the
sample. A third is that there could have been a change
in the cosmic ray flux {which on this site is estimated
to provide approximately half the external dose-rate)
due to a change in the degree of shielding by rock.
lf leaching had occurred a progressive increase in
external dose-rate with depth would be expected.
Reference to Table 1 ofVa1ladas et al. (1988) clearly
shows this not to be the case: in fact the extreme
values of 0.22 and 0.25 Gy ka'1 are easily contained
within the quoted area limits (of; 0-04 Gy ka`l).
Further, ifleaching had occurred it would be expected
that there would be disequilibrium in the uranium
series; but measurements using ot and Y spectrometry
showed this not to be the case.
There is no direct evidence bearing on the other
two possibilities but there is implicit indication in the
reported data that any past variations in external
dose-rate (and indeed in internal dose-rate too) have
not been serious enough to distort the ages obtained.
The internal dose-rates of` the individual Hints are
widely spread: from 0.41 to 1.36 Gy ka'1, unrelated to
depth. Yet the ages obtained for Hints of high internal
dose-rate are not significantly different to those
obtained for Hints of low internal dose-rate (see Hgure
3a); if there was a significant systematic error in either
dose-rate evaluation this would not be the case.
Similar evidence of reliability comes from isochron
analysis, first used in TL dating by Mejdahl (1983) and
demonstrated for ESR by Blackwell & Schwarcz
(1992). This is possible because of the combined
circumstances of a wide spread in the internal dose
rates, constancy of external dose-rate down the sec
tion, and the geological indications (see Valladas et al.
1988) that the sediment of the section accumulated
Phil. Trans. R. S0:. Lond. B (1992)
rapidly. Isochron analysis allows evaluation of aver
age age without insertion of any value for the external
dose-rate; it is only required that this latter is the same
for all samples. Figure 3b shows its application: the
age so evaluated is 88 (39) ka; the published age
(92 3 5) is consistent with this.
ESR dates for tooth enamel from the burial levels
have also been obtained (Schwarcz et al. 1988) giving
an early uptake (EU) age of 96 (3 13) ka and a linear
uptake age (LU) of 115 (3 15) ka, the latter being
favoured. These dates utilized the external dose-rate
that was evaluated for the TL but in the case of ESR this
is the dominant contribution to the overall closerate.
Hence the two techniques are not strongly interdepen
dent as far as dose-rate is concerned, and given the
different nature of the crystal structures in which the
trapped electrons were accumulated the ESR result can
be considered as independent conHrmation of the sites
great age. The fact that the favoured ESR age of
115 (3 15) ka is somewhat greater than the TL age
(9235ka) may be due to underestimation of the
external dose-rate. The isochron analysis of the TL
dates (figure 3b) indicates an external dose-rate of
0.28 (30.09) Gy ka-] whereas the values used (for
both TL and ESR) were in the range 0.22-0.25 Gy ka`
Because of the dominance the external dose-rate in the
case of EsR, use of the isochron value would tend to
move the ESR result into agreement with the TL.
Finally, it is to be noted that the scatter in the
twenty individual TL ages corresponds to a standard
deviation of only 8%; this is very close to the predicted
random (Statistical) error limits quoted for the
individual dates. Hence these twenty ages form a
remarkably coherent group, thus giving further strong
support to the validity of the result. Such coherence is
not alwavs the case: for instance the six individual TL
ages on which the average of 119 ka for es-Skhul (see
2a) are based show a much wider scatter, the
standard deviation being around 18%. This is indica
tive of less favourable circumstances, as was recog
nized by Mercier el al. (1992) who quoted predicted
error limits of 3 18 ka. Thus in making archaeological
interpretation of these results for the presence in the
region of anatomically modern humans, it is appro
priate to put emphasis on the TL result for Qafzeh.
5. SUMMING UP
Luminescence techniques allow the independent dat
ing, directly in calendar years, of burnt Hint, burnt
quartz and calcite, as well as unburnt wind- and
water-borne sediment; for the latter it is advantageous
to use the osL technique. Consideration of the burnt
Hint TL data for Qafzeh underlines the reliability of
that result, one of several which are critical for our
understanding of the origin of modern humans.
We are grateful to M. Whitehouse for carrying out the
isochron plot and to S. G. E. Bowman, E. Rhodes, H. P.
Schwarcz and S. Stokes for helpful comments. OCR Output
144 M. Aitken and H. Valladas Luminescence dating
REFERENCES
Aitken, MJ. 1985 Thermaluminexcence dating. London: Aca
demic Press.
Aitken, MJ. 1989 Luminescence dating: a guide for non
specialists. Archaeametry 31, 1474159.
Aitken, 1990 Science-based dating in archaeology. London
& New York; Longman.
Aitken, MJ. 1992 Optical Dating. Quat. Sci. Rez;. 11, 127
132.
Aitken, MJ. & Bussell. G.D. 1982 TL dating of fallen
stalactites. PACT 6, 550554.
Bai1if LK. (ed.) Ancient TL Date Lists. Lumineseence
Dating Laboratory, University of Durham.
Bard, E., Hamelin, B., Fairbanks, R.G. & Zindler, A. 1990
Calibration of the C-14 timescale over the past 30,000
years using mass-spectromctric U-Th ages from Barbados
corals. Nalure, Lond. 345, 405-410.
Bell, W.'I`. 1991 Thermoluminesccnce dates for the Lake
Mungo Aboriginal Hreplaces and the implications for
radiocarbon dating. Arclzaeometry 33, 43-50.
Berger, G.\/V. 1988 Dating Quaternary events by lumines
cence. Geo!. Soc. Am. spec. Pap. 227, 13-50.
Blackwell, B. & Schwarcz, H.P. 1992 Electron spin
resonance isochron dating. Appl. Rad. Isotopes (In the
p1"CSS.)>
Bowman, S.G.E. 1985 Thermoluminescence characteristics
of sediments from the Tabun cave, Israel. Nucl. Trac/rs
Rad. Meas. 10, 731-736.
Debenham, NC. & Aitken, MJ. l984 Thermolumines
cenee dating of stalagmitie calcite. Arclzaeometry 26, 155
170.
Griin, R., Stringer, CJ. 8; Schwarcz, H.P. 1991 ESR
dating of teeth from Garrods Tabun cave collection. J.
hum. Evol. 20, 231-248.
Huntley, DJ., Godfrey-Smith, D.I. & Thewalt, M.L.\/V.
1985 Optical dating ofsedimcnts. Nature, Land. 313, 105
107.
Huxtable, & Aitken, MJ. 1977 Thermoluminescence
dating of Lake Mungo polarity excursion. Nature, Lond.
265, 40-41.
Huxtable, & Aitken, MJ. 1985 Thermoluminescence
dating results for the Palaeolithic site MastrichtBe1v
dere. Analecta Praehzszorica Leidenria 18, 41-44.
Huxtable,_]. & Aitken, MJ. 1988a Datation par thermolu
minescence. In Le gisement palolit/zique moyen de Biache
Saint- Vaast (Pas de Calais }: stratigrap/tie, environment, tudex
archalugiquee, vol. 21 (ed. A. Tuffreau & Somm), pp.
107-108. Mmoire de la socit prhistorique francais.
Huxtable,J. & Aitken, 1988b Datation par thermolu
minescencc de la grotte Vaufrey. In La gratte Vaufrey d
Cnac et Saint Julien (Dordogne): paloenvironmerzts, chronolgie,
et activites humaines, vol. 19 (ed. P. Rigaud), pp. 359
364. Mmoire de la socit prhistorique francais.
Mazaud, A., Laj, C., Bard, E., Arnold, M. & Tric, E. 1991
Geomagnetic field control of C-14 production over the
last 80 ky: implications for the radiocarbon timescale.
Geophys. Res. Lett. 18, 1885-1888.
Phil. Trans. R. Soc. Land. B (1992)
McKeever, S.W.S. 1985 Thermoluminescence ty solids. Cam
bridge University Press.
Mejdahl, V. 1983 Feldspar inclusion dating of ceramics
and burnt stones. PACT 9, 351-364.
Mejdahl, V. 1988 The plateau method for dating partially
bleached sediments by thermoluminescence. Qaat. Sci.
Rev. 7, 347-348.
Melcher, C.L. & Zimmerman, D.W. 1977 Thermolumi-
nescent determination of prehistoric heat treatment of
chert artefacts. Science, Wash. 197, 1359-1360.
Mercier, N. 1992 Apport des mthodes radionuclaires
de datation A 1tude du peuplement de lEurope et du
Proche-Orient du Pleistocene moyen et suprieur. These
prsente A 1Universit de Bordeaux 1, 139 pp.
Mercier, N., Valladas, H., Bar-Yosef; O., Vandermeersch,
B., Stringer, C. &_]oron,J-L. 1992 Thermoluminescence
date for the Mousterian burial site of es-Skhul, Mt.
Carmel. J. archaeol. Sci. (In the press.]
Roberts, R.G., Rhysjones 8; Smith, M.A. 1990 Thermolu
minescence dating of a 50,000-yearold human occupa
tion site in northern Australia. Nature, Lond. 345, 153-156.
Robins, G.V., Seeley, NJ., Symons, M.C.R. & Bowman,
S.G.E. 1982 The Institute of Archaeology (London)
ESR flint project. PACT 6, 322-332.
Schwarcz, H.P., Grun, R., Vandermeerch, B., Bar-Yosef,
O., Valladas, H. & Tchernov, E. 1988 ESR dates from
the burial site of Qafzeh in Israel. J. hum. Evol. 17, 733
737.
Smith, B.W., Rhodes, EJ., Stokes, SJ., Spooner, N.A. &
Aitken, MJ. 1990 Optical dating of sediments: initial
quartz results from Oxford. Archaeometry 32, 19-31.
Stokes, S. & Gaylord, D.R. 1992 Geochronology of the
Cleer Creek dune field: a test of optical dating. Quat. Res.
(In the press.)
Stokes, S. 1993 Optical dating of sediment samples from
Bir Tafawi and Bir Sahara: an initial report. In The
Middle Palaeolithic in the Egyptian Sahara during the last
interglacial (ed. F. WendorlQ A. E. Close & S. C. Schilde).
Southern Methodist University Press. (In the press.;
Valladas, H. 1992 Thermoluminescence dating of flint.
Qual. Sci. Rev. ll, 1-5.
Valladas, H. Joron, _].L., Valladas, G., Arensburg, B., Bar
Yosei; O., Belfer-Cohen, A., Goldberg, P., Laville, 1-1.,
Mcignen, L., Rak, Y., Tchernov, E., Tiller, A.M. &
Vandermeersch, B. 1987 Thermoluminescence dates for
the Neanderthal burial site at Kcbara in Israel. Nature,
Lond. 330. 159-160.
Valladas, H, Reyss, _].L., Joron, _].L., Valladas, G., Bar
Yoseh O. & Vandermersch, B. 1988 Thermolumines
cence dating of Mousterian Prot0-Cro-Magnonl remains
from Israel and the origin of modern man. Nature, Lond.
331, 614-616.
Wintle, A.G. 1990 A review of current research on TL
dating of loess. Quat. Sci. Rev. 9, 385-397.
York, D. 1969 Least squares fitting with correlated errors.
Earth planet. Sci. Lett. 5, 320-324.
Zoller, L. & Wagner, G.A. 1990 Thermoluminescence
dating of loess. Quat. Int. 7/8, 119-128. OCR Output
Outlining thc problem
P. A. MELLARSK M. AITKENZ AND C. B. STRINGER
I Department yArc}zae0l0gy, University J Cambridge, Downing Street, Cambridge CB2 3DZ, U.K.
2 Rexearc/2 Laboratory for Archaeology, Oxford University, 6 Keble Road, Orgrord OX] 3QJ, U.K.
3 Human Origins Group, Department of Palaeontology, Natural History Museum, London S IV 7 5BD, UK.
Few ropics have generated more debate and contro
versy in the scientific literature over the past few years
than the biological and behavioural origins of anato
mically modern human populations: i.e. populations
belonging to our own form of Homo sapiens sapiens.
What is common ground in all these debates is that
populations that were fundamentally `modern in both
a basic anatomical sense, and in at least the majority
of cultural and behavioural senses, were eH`eetively
established throughout all the major regions of the
Old \Vorld [i.e. Africa, Asia. Europe and parts of
Australasia) by at least 30-35 ka ago (1ka=l000
years). The current controversy hinges essentially on
what happened before this time above all during the
crucial formative phases of modern human develop
ment between ca. 200 and 30 ka ago. It is this period
which forms the major battle ground ofcurrent debate
over the evolutionary origins and mutual relationships
of modern world populations, and which provides the
central focus ofthe present symposium.
Throughout the greater part ofthe present century
effectively since the discovery ofthe `classic Nean
derthal skeleton at La Chapelleaux-Saints in 1908
the issue ofthe biological and demographic origin of
modern human populations has been seen very largely
as a dichotomy between two sharply opposed schools
of thought. One view- frequently referred to in the
more popular scientific literature as the Garden of
Eden or Noahs Ark hypothesis asserts that biolo
gically and genetically modern human populations
evolved initially in one fairly limited and closely
prescribed region of the world and subsequently
dispersed at varying times to all other regions,
either with or without signihcant degrees of genetic
intermixture and interbreeding with the pre-existing
populations within the same regions. At various points
throughout the present century, and with varying
degrees of emphasis, this view has been maintained by
{among others) Marcellin Boule, \\illiam Howells and
F. Clark Howell. More recently, workers such as
Giinter Brauer, as well as one ofthe present authors
fe.g. Stringer & Andrews l988), have focused atten
tion on Africa as the particular continent of origin: the
so-called Out of Africa model.
The principal alternative scenario- now generally
referred to as the muItiregiona1 evolution hypothe
sis - asserts that there was no such simple or discrete
origin for the morphological and genetic character
istics of modern humans, and that human populations
in all of the major regions ofthe Old World (i.e.
Africa, Asia and Europe) have pursued an essentially
Phil. Trans. R. Soc. Lond. B {1992) 337, 127-130
. Printed in Great Britain OCR Output
parallel pattern of evolution over at least the greater
part of the past million years or so. Central to the
latter model, of course, is the assumption that there
was a sufficient degree of interbreeding and resulting
gene flow between these geographically dispersed
populations to maintain a broadly similar pattern of
evolutionary development within the different regions
throughout the whole of this time range. The princi
pal proponents of this hypothesis have been Ales
Hrdlicka, Franz Weidenreich, Carleton Coon, Loring
Brace and (more recently) Milford Wlolpoff and Alan
Thorne (e.g. Wolpoff 1989; Thorne & WolpoH` 1992).
Potentially one of the most significant points of
conflict between these two models relates to the
character and timing ofthe divergence of the modern
patterns of regional (racial) variation among pres
ent-day populations. \/Vhereas the population-disper
sal hypothesis asserts that all of these divergences are
the result of relatively short-term evolutionary pro
cesses operating purely over the past l00 ka or so, the
multiregional evolution viewpoint would trace these
regional divergences back through a much longer
timescale of at least several hundred thousand to one
million years.
It is hardly possible here to review the ebb and flow
of debate between these two condicting hypotheses
over the past 80 years or so. Reviews ofthe earlier
years of this debate have been provided by Frank
Spencer (1984), and more recently by Smith et at.
(1989). \iVhat can be said is that over the course ofthe
past 10-15 years, a number of significant develop
ments emerged which appeared to shift the balance of
the evidence fairly strongly in favour of the former
(population dispersal) scenario. One of these develop
ments was the application ofnew dating methods. The
use ofthe newly developed dating methods of thermo
luminescence {TL) to burnt flint and electron spin
resonance (esa) to tooth enamel associated with the
large samples of skeletal remains recovered from the
two sites of Mugharet es Skhul and jebel Qafzeh in
northern Israel revealed that these fossils - which had
previously been generally assumed to be around 40
50 ka in age - in reality dated to around 100 ka (Griin
& Stringer 1991; Aitken & Valladas, this symposium;
Schwarcz & Griin, this symposium). Because these
fossils had generally been accepted and described in
the literature as showing modern features, this
appeared to demonstrate that such forms had
appeared in at least one region ofthe world at a much
earlier date than suspected hitherto, and demon
strablv at a time when much more archaic forms
127 1992 The Royal Society and the authors
128 P. /\. Mellurx and others Oullining the problem
Qsuch ns thc Nvz111dcrtl1aIs) were still living in closely
z1d_izuen1t or even overlapping- areas of Eurasia.
l`ur1l1er discoveries seemed to reinforce the same
pntterii. Thus a number of` other finds of essentially
`moderif skeletal remains at sites in southern and
easrerii Africa (such as Border Cave and Klasies River
Mouth in South Africa, and Omo in Ethiopia)
appeared to date from around 80-100 ka ago
X;Rightmire l989), whereas, by contrast, a typically
Neanderthal skeleton from the site of SaintCsaire
in western France was found to date from as
recently as ca. 35 ka ago (Lvque & Vandermeersch
1980; Mercier ez al. 1991). All of these discoveries
appeared, on the face of it, to conform much better
with the predictions of the population dispersal
scenario of modern human origins, than with the
alternative hypothesis of` essentially parallel, multi
regional evolution within the different regions of the
wOrld.
The final, and perhaps most significant discovery
which was claimed to argue specifically in favour of
the population dispersal hypothesis emerged from the
detailed work carried out on the patterns of mito
chrondrial DNA variation in modern human popu
lations in different regions of the world, undertaken
by the late Allan C. Wilson and his colleagues at the
University of California at Berkeley (eg. Cann el al.
1987; Stoneking & Cann 1989; VVilson & Cann 1992).
Basically, \/Vilson and his co-workers argued that
modern human populations were far too similar in
terms of their mitochondrial DNA makeup to be the
products of largely independent evolution over a span
ofa million years or so, and must be the products of
divergence from a single common ancestor at a much
more recent period. Arguing that variations in
mtDNA were most probably neutral in an adaptive
sense and that the rate of mutation of mtDNA can be
estimated from studies in other animal groups, they
postulated that a common female ancestor had prob
ably lived in the region of 200 ka ago (more broadly
between ca. 50 and 350 ka), and that the descendants
of this initial population had probably expanded to
most other regions of the world by around 30-100 ka
ago. Finally, they suggested that the geographical
source of this common ancestor was most probably
located somewhere in Africa. The latter conclusion
was based partly on the demonstrably wider range of
genetic variability apparent in modern African popu
lations than among those in other parts of the world
(suggesting a longer period of evolution of modern
humans in this region than elsewhere) and partly on
the construction of hypothetical trees of genetic de
scent by means of maximum parsimony methods, all
of which seemed to point to an African origin. At
about the same time, similar claims for an apparently
African origin were made by a number of workers
based on analogous studies of the patterns of variation
in classical genetic markers, such as blood groups
and proteins (e.g. Cavalli-Sforza et al. 1988), and by
other workers based on variations in nuclear DNA in
modern populations (e.g. \/Vainscoat ez al. 1989;
Lucotte 1989; Cavalli-Sforza 1991; Mountain et al.,
this symposium). One of the most critical and contro
Phil. Tran:. R. Sec. Lond. B (1992)
versial conclusions of the mtDNA studies was that
the process of hypothetical population dispersal
from the presumed African homeland appeared to
have been achieved with no detectable interbreeding
with the earlier, biologically archaic populations
within the different regions of Eurasia, and that
these populations may therefore have become _
extinct without contributing any significant genetic
legacy to the subsequent populations in the same
regions (Stoneking & Cann 1989; Willson & Cann
1992).
Reactions to these claims by proponents of the
multiregional evolution school were swift, pointed,
and forcefully argued (e.g. Wlolpoff 1989; Thorne &
\No1poff 1992). Several workers pointed to the severe
problems of estimating the rates of genetic divergence
of mtDNA, and argued that by adopting rather
different mutation rates (based, for example, on the
assumption of an earlier date of divergence between
the chimpanzee and human lineages, or earlier dates
for the human colonization of the Americas and
Australasia) one could recalibrate the date of the
inferred common ancestor of all modern populations
closer to one million years: that is, near the generally
accepted date for the initial colonization of northern
latitudes and eastern Asia by Homo erertus populations
in the early Pleistocene. Further debate has centred on
the methods of constructing the hypothetical genetic
treesparticular1y those involving the assumptions
and methodology of maximum parsimony
approaches a debate which fiared up once again in
the literature in the immediate run-up to the present
symposium (e.g. Nladdison 1991; Templeton 1992;
Stoneking el al., this symposium). If these objections
are valid, then the presumed common ancestor of
modern populations could potentially be located in
almost any part of the Old \Vorld, rather than just
Africa. Similarly, the implications of the skeletal
remains themselves have become a topic of lively
debate. 1/Volpoff, Thorne and others have argued that
many of the current interpretations of individual
skeletal remains (such as the recently discovered
Neanderthal from SaintCsaire) fail to make suf
ficient allowance for the probable degree of intrapo
pulation variability in Pleistocene hominids. More
specifically, they have argued that the morphology of
some of the relatively recent hominid remains from
areas such as southeast Asia, Australasia and Central
Europe, point strongly to a substantial element of
morphological and genetic continuity between the
archaic and anatomically modern populations in
these regions (e.g. WolpoH` 1989; Thorne & W'olpoff
1992).
The spate of debate currently surrounding all of
these issues in the origins and evolution of modern
human populations served as a primary catalyst for
the organization of the present meeting. The aim,
essentially, has been to bring together a range of
specialists in relevant scientific disciplines (i.e. evolu
tionary anatomy, molecular biology, archaeology,
dating methods, etc.) to review the most recent
discoveries and changing theoretical perspectives in
these wide-ranging debates. The central theme of the OCR Output
 Outlining the problem P. A. Mellars and others 129

meeting ofcourse was to assess the specifnc bearing of
recent advances in various scientinc dating methods
0n the interpretation of these crucial formative phases
in modern human evolution. lt is this Field of research
which has made some of the most striking advancesr
over the past decade, and which as indicated in the
preceding discussion is now proving to be critical in
the systematic testing ofthe various alternative models
for the emergence of modern human populations.
The chronometric techniques primarily relevant to
the topic of this volume (i.e. uranium-series, lumi
nescence, electron spin resonance and amino acid
racemization] provide dates for such materials as
stalagmitic calcite, burnt flint, sediment, tooth enamel,
and ostrich eggshells. Most of these techniques have
reached the stage of application to the Palaeolithic
comparatively recently, and the results so far achieved
are only a foretaste of the full realization of their
potential. The longerestablished radiocarbon method
unfortunately has an effective limit of around 40 ka
(except in special circumstances) and hence only
reaches into the later phases of modern human
development, for which its datings of bone, teeth and
charcoal are indeed of critical importance. At the
other end of the timescale, in the dating of early
hominids, the potassium-argon and fission-track tech
niques have been of even greater importance.
Although the time-range ofthe former now extends
into the Holocene, this method has not vet had
significant impact on the issues under discussion here.
This is because it is limited to circumstances in which
there is chance association of human occupation with
volcanic products (e.g. lava Hows, air~fall deposits).
Fuller discussion of these and other chronometric
techniques will be found elsewhere (eg. Aitken 1990).
Another important aspect of Palaeolithic chrono
logy is the climatic framework within which sites are
placed. Sites can be attributed to such climatic
frameworks on several bases: (i) on their fossil flora,
fauna and geological features; (ii) the characteristics of
the associated archaeological assemblages; or, more
recently, (iii) by direct chronometric dating. Initially,
the climatic framework was in terms of the four major
glaciations {Giinz, Mindel, Riss and Wurm) that had
been recognized on the basis of geological evidence in
the Alps; more complex regional frameworks were
later evolved for example, inclusion of warmish
interstadials within the long glacial periods with a
correspondingly complex nomenclature. Fortunately a
worldwide framework is now available. This is the
oxygen-isotope stratigraphy, based on the climatic
dependence of the ratio ISO/ISO in marine fossil
microfauna, studied in long cores obtained from the
ocean floor; this ratio is used to define a succession of
alternating cool and warm stages. The remanent
magnetization of the terrigeneous component of the
sediment cores allowed correlation of one stage with
the most recent major geomagnetic polarity reversal
observed in volcanic rocks, for which absolute dating
has been provided by the potassium-argon technique;
other stages were dated by assuming a constant rate of
sedimentation, with further control provided by radio
carbon and uranium-series datings.
Latterly it has been established that there is good

IW!} 3

I1 Il
|| clle
I I
lbl _l e

100 200 300

time / ka BP

Figure l. Oxygen-isotope variation for the past 300 ka with astronomically based timescale (redrawn, with
additions, from Martinson et al. (1987)). The vertical axis represents changes in the averaged ISO/160 ratio found in
benthic (bottom-living) foraminifera from Eve locations in the oceans of the world; Eve divisions equal a change of
one part per thousand. Along the top the numbers allocated to warm stages are given, with intervening (even
numbered) cold stages being shown shaded. The letters (a, c, e) refer to warm substages of stage 5; there are
intervening cool troughs, b and d.

Phi!. Trans. R. S0:. Lond. B (1992) OCR Output

130 P. A. Mcllars and others Outlining the problem
correlation between the pattern of climatic variation
so revealed and the Milankovitch astronomical pre
dictions based on changes in the earths orbital motion
(eccentricity ofthe orbit around the sun; obliquity of
the ecliptic; procession of the equinoxes); the changes
result from gravitational perturbations due to the
changing conhguration of the planets. VVith this
correlation established, the oxygenisotope stages
became datable with the same high accuracy as is
possible for the orbital changes. In the context ofthe
present symposium, the chronostratigraphy developed
for the past 300ka by Martinson el al. (1987) is of
particular relevance. According to this (see figure l)
the penultimate major glaciation (stage 6) ended
about 130 ka ago and the succeeding interglacial
complex (stage 5) lasted until around 74 ka ago. The
degree to which there is exact land-ocean synchro
neitv for all the detail revealed in the ocean sediments
is the subject of continuing research, but the main
features seem to be well reproduced on the continental
land masses, thus giving a general chronological
framework that is well dated in absolute terms. The
isotope stages are now the fundamental time divisions
ofthe Quaternary epoch.
Finally, it needs to be realized that although the
new chronometric techniques mentioned earlier may
be having a strong impact, there can be site-to-site
variability in the reliability of results, and also that
some sample types are not as satisfactory as others.
Also, as some dating applications are made with the
objective of exploring the performance ofa technique
in particular circumstances, it is important for a
consumer to be aware of the pedigree ofa date and to
use restraint, if appropriate, in its utilization. For such
discretion to be exercized it is necessary that the date
should have been presented for publication in sufH
cient detail for assessment ofits reliability to be made,
at any rate by a specialist colleague; dates not so
presented should be treated with caution, or ignored.
Obviously it is advantageous if the consumer has
familiarity with the dating method concerned, and
one objective of the meeting on which this symposium
is based was to encourage dialogue between consumer
and dating specialist; hence the dating contributions
which follow are addressed primarily to the non
specialist.
Note on dating terminology
The letters BP signify before present. They have the
strict connotation that the age quoted is given in
radiocarbon years rather than calendar years; radio
carbon ages that have been converted into calendar
years by calibration are given as cal BP. Most other
techniques yield ages directly in calendar (i.e. side
real) years and the letters BP are then inappropriate;
in the case of amino acid dates based on calibration
against radiocarbon, BP should be retained.
P/zil. Trans. R. Sac. Lond. B (1992)
REFERENCES
Aitken, M,]. 1990 Science-based dating in archaeology. London
& New York: Longman.
Cann, R.L., Stoneking, M. & Wilson, A.C. 1987 Mito
chondrial DNA and human evolution. Nature, Lond. 325,
31-36.
Cavalli-Sforza, L.L. 1991 Genes, peoples and languages.
Scienl. Arn. 265(1l) (November 1991), 72-78.
Cavalli-Sforza, L.L., Piazza, A., 1\{enozzi, P. & Mountain,
_]. 1988 Reconstruction of human evolution: bringing
together genetic, archeological and linguistic data. Proc.
naln. Acad. Sci. U.S.A. 85, 8002-8006.
Griin, R. & Stringer, C.B. 1991 Electron spin resonance
dating and the evolution of modern humans. Archaeometry
33, 153-199.
Lvque, F. & Vandermeersch, B. 1980 Dcouverte des
TCSICS humains dans un niveau castelperronien a Saint
Csaire (Charente Maritime). C. r. Acad. Sci., Paris (Series
II) 291, 187-189.
Lucotte, G. 1989 Evidence for the paternal ancestry of
modern humans: evidence from a Y-chromosome specilic
sequence polymorphic DNA probe. ln l\rIellars & Stringer
(1989), 39-46.
Maddison, D.R. 1991 African origin of human mitochon
drial DNA reexamined. Syst. Zool. 40, 355-363.
Martinson, D.G., Pisias, N.G., Hays, _l.D., lmbrie, _].,
Moore, T.C. & Shackleton, 1987 Age dating and
the orbital theory of the ice ages: development of a high
resolution O to 300,000-year chronostratigraphy. Qual.
Res. 27. 1-29.
Mellars, P. & Stringer, C. (eds) 1989 The human revolutio: ;
behavioural and biological perspectives on lhe origins ry modern
humans. Edinburgh University Press.
Mercier, N., Valladas, H.,_]oron,_].L., Reyss,_].L., Levque,
F. & Vandermeersch, B. 1991 Thermoluminescence
dating of the late Neanderthal remains from Saint
Csaire. Nature. Lond. 351. 737-739.
Rightmire, G.P. 1989 Middle Stone Age humans from
eastern and southern Africa. In Mellars & Stringer
(1989), 109-122.
Smith, F.H., Falsetti, A.B. & Donnelly, S. 1989 Modern
human origins. Yb. phys. Anl/zrop. 32, 35-68.
Spencer, F. 1984 The Neanderthals and their evolutionary
significance: a brief historical survey. In The origins of
modern humans: a world survey ty the jizssil evidence (ed. F. H. -4
Smith & F. Spencer), pp. 1-49. New York: Alan R. Liss.
Stoneking, M. & Cann, R.L. 1989 African origin of human
mitochondrial DNA. In Mellars & Stringer (1989), 17
30.
Stringer, C.B. & Andrews, P. 1988 Genetic and fossil
evidence for the origin of modern humans. Science, l/Vash.
239, 1263-1268.
Thome, A.G. & 1/Volpofli M.H. 1992 The multiregional
evolution of humans. Scient. Am. 266(4) (April 1992), 28
33.
Wlainscoat, _].S., Hill, A.V.S., Thein, S.L., Flint, _]., Chap
man, _].C., Weatherall, DJ., Clegg, _].B. & Higgs, D.R.
1989 Geographic distribution of alpha- and beta-globin
gene cluster polymorphisms. In Mellars & Stringer
(1989), 39-46.
Wilson, A.C. & Cann, R.L. 1992 The recent African
genesis of humans. Scient. Am. 266(4) (April 1992), 22-27.
Wolpoff, M.1-1. 1989 Multiregional evolution: the fossil
alternative to Eden. In Mellars & Stringer (1989), 62
108.