U-Series Dating of Paleolithic Artin 11 Caves in SpainA. W. G. Pike,1* D. L. Hoffmann,2,3 M. García-Diez,4 P. B. Pettitt,5 J. Alcolea,6 R. De Balbín,6

C. González-Sainz,7 C. de las Heras,8 J. A. Lasheras,8 R. Montes,8 J. Zilhão9

Paleolithic cave art is an exceptional archive of early human symbolic behavior, but becauseobtaining reliable dates has been difficult, its chronology is still poorly understood after morethan a century of study. We present uranium-series disequilibrium dates of calcite depositsoverlying or underlying art found in 11 caves, including the United Nations Educational, Scientific,and Cultural Organization (UNESCO) World Heritage sites of Altamira, El Castillo, and Tito Bustillo,Spain. The results demonstrate that the tradition of decorating caves extends back at leastto the Early Aurignacian period, with minimum ages of 40.8 thousand years for a red disk,37.3 thousand years for a hand stencil, and 35.6 thousand years for a claviform-like symbol.These minimum ages reveal either that cave art was a part of the cultural repertoire of the firstanatomically modern humans in Europe or that perhaps Neandertals also engaged in painting caves.

European Upper Paleolithic cave painting

and engraving are among some of the

earliest examples of art and human sym-

bolic behavior, although there is considerable

uncertainty in when they began and how styles

and practices developed. Accurate dating would

help determine whether they arrived with the

earliest populations of anatomically modern hu-

mans by 35 to 40 thousand years ago, were a by-

product of their interaction with Neandertals, or

developed later (1–4). Distinct phases are recog-

nized in the art, based on technique, theme, style,

and superimposition (5), but it is unclear whether

these overlapped or evolved. Engravings and,

in many cases, paintings lack organic pigments

or binders suitable for accelerator mass spectro-

metry radiocarbon dating (6). Where suitable

material exists (e.g., charcoal pigments), only

small samples can be dated so as to minimize

damage to the art, magnifying the effects of con-

tamination and resulting in larger uncertainties.

Discrepancies between multiple 14C determina-

tions on a single painted motif have been com-

mon, as are discrepancies between the dates of

different chemical (e.g., humic/humin) fractions

of the same sample (7).

We used uranium-series disequilibrium to date

the formation of thin calcite flowstone growths

that formed on the surfaces of paintings and en-

gravings in 11 caves in Asturias and Cantabria,

northwestern Spain. This approach circumvents

the problems related with radiocarbon dating (6)

and yields minimum ages for the art. In some cases

where paint has been applied atop a flowstone

or a flowstone has been engraved, maximum

ages can also be obtained (8). Uranium-series iso-

topes can be measured in samples as small as

10 mg (9, 10), which makes it possible to acquire

samples whose stratigraphic relationship with the

art is robust.

Results

We obtained 50 calcite samples that overlay

paintings or engravings from the 11 cave sites in

Asturias andCantabria in northwest Spain (Fig. 1).

Samples of between 10 and 150 mg were ex-

tracted by scraping with a blade or with an

electric drill. We sampled calcite covering a vari-

ety of art and representing a range of styles. The

samples were processed and U-series isotopes

measured by using the method of Hoffmann et al.

(9–11).Where sampling allowed a second aliquot

to be taken, we tested the integrity of the calcite

by comparing the dates of the upper layers of the

calcite to those closer to the painting. In all cases,

the date from the deeper sample was older, sup-

porting the reliability of our method (11).

Our U-series ages ranged from 0.164 to 40.8 ky

[corresponding to radiocarbon ages of near mod-

ern to 35,500 radiocarbon years before the present

(14C yr B.P.)]. Figure 2 shows calculated uranium

series dates, from calcite on top of paintings, ar-

ranged in ascending order (also table S1). A paint-

ing cannot be younger than the calcite that formed

on top of it, but there is an unknown delay be-

tween the execution of the painting and the for-

mation of the calcite, so we present the data as the

cumulative proportion, p, of paintings we have

dated that cannot be younger than the date, T

(Fig. 2 and table S1).

For context in discussing the ages of the

paintings, the earliest reported 14C date for the

Proto-Aurignacian culture in northern Spain, as-

sumed to represent the arrival of Homo sapiens,

is from the site of Morin at 36,590 T 770 14C yr

B.P., yielding (12) a 2-s calibrated age of 40,037

to 42,778 calendar years before the present (cal yr

B.P.). This has been thought to be followed by

the Aurignacian I at about 40,000 cal yr B.P.

and the Aurignacian II culture at about 37,000

cal yr B.P. (13). In France and Iberia, the latest

Aurignacian dates from 37,000 to 34,500 cal yr

B.P. (14). In Cantabria, the latest Aurignacian is

possibly represented at Cuco and dated to 30,020 T

160 14C yr B.P. (34,490 to 35,032 cal yr B.P.)

(15). The earliest Gravettian levels in northern

Spain are around 27,400 14CyrB.P. at Antoliñako

and Almada; the most precise age is from

Antoliñako at 27,390 T 320 14C yr B.P. or 31,145

to 32,487 cal yr B.P. (16). The latest Gravettian is

represented by a date of 20,124 T 340 14C yr B.P.

(23,275 to 24,975 cal yr B.P.), again from Morin

(16), which is within error of the earliest date on

Solutrean levels fromLaRiera (20,970 T 620 14C

yr B.P., 23,578 to 26,921 cal yr B.P.). In light of

other ages from southwest Europe (17, 18), the

RESEARCHARTICLE

1Department of Archaeology and Anthropology, University ofBristol, 43 Woodland Road, Bristol BS8 1UU, UK. 2BristolIsotope Group, School of Geographical Sciences, University ofBristol, University Road, Bristol BS8 1SS, UK. 3Centro Nacionalde Investigatión sobre la Evolución Humana (CENIEH), PaseoSierra de Atapuerca s/n, 09002 Burgos, Spain. 4Department ofGeography, Prehistory and Archaeology, University of BasqueCountry (UPV/EHU), c/ Tomás y Valiente s/n, 01006 Vitoria,Spain. 5Department of Archaeology, University of Sheffield,NorthgateHouse,West Street, Sheffield S1 4ET, UK. 6PrehistorySection, University of Alcaláde Henares, c/ Colegios 2, 28801Alcaláde Henares, Madrid, Spain. 7Department of Historic Sci-ences, University of Cantabria, Avenida Los Castros s/n, 39005Santander, Spain. 8MuseoNacional y Centro de Investigación deAltamira. 39330 Santillana del Mar, Cantabria, Spain. 9Uni-versity of Barcelona/Institució Catalana de Reserca i EstudisAvançats (ICREA), Departament de Prehistòria, Història Antiga iArqueologia (SERP), c/ Montalegre 6, 08001 Barcelona, Spain.

*To whom correspondence should be addressed. E-mail:alistair.pike@bristol.ac.uk

Fig. 1. Locations of the cavessampled: 1, Pedroses; 2, TitoBustillo; 3, Las Aguas; 4, Altamira;5, Santiàn and El Pendo; 6, ElCastillo, La Pasiega, and LasChimeneas; 7, Covalanas andLa Haza.

www.sciencemag.org SCIENCE VOL 336 15 JUNE 2012 1409

on J

une 1

4,

2012

ww

w.s

cie

ncem

ag.o

rgD

ow

nlo

aded fro

m

latter date is a more reliable indicator of the

timing of the Gravettian-Solutrean transition in

northern Spain.

Nearly 20% of our dates on cave paintings

fall into pre-Magdalenian times (i.e., before

21,000 cal yr B.P.), and of these four are pre-

Gravettian. This distribution is not meant to rep-

resent relative intensity of artistic activity over

time because of sampling bias, including those

caused by cave settings and the influence of cli-

mate on calcite growth (19), but the dates indicate

that early painting was not a one-off activity. Be-

low and in Table 1 and Fig. 3, we focus on several

of the more interesting (pre-Magdalenian) dates.

Altamira. Altamira cave, on the northern

coast of Spain, contains numerous paintings, in-

cluding of human hands and animals. The chro-

nology of the art has been debated since its

discovery, particularly since Breuil developed

a scheme based on the superimpositions of

several paintings (20, 21). Although there is gen-

eral agreement that several phases are recog-

nizable in the cave’s 10 main decorated zones,

these have been suggested to span either a rela-

tively short period of time that corresponds broad-

ly to the cave’s known archeology—that is, the

Solutrean and the Magdalenian, Leroi-Gourhan’s

styles III and IV (22, 23)—or considerably

longer (21). We obtained a minimum age of

22.0 thousand years (ky) for the red dot outline

horse on the ceiling of the polychrome cham-

ber (O-53); thus, it is at least Solutrean in age.

A second minimum age of 35.6 ky was obtained

for the large claviform-like symbol (O-50) in the

Table 1. Results of U-series disequilibrium dating for samples mentioned inthe text. All isotopic ratios are activity ratios; errors are at 2s. Ages arecorrected for detritus by using an assumed 232Th/238U activity of 1.250 T

0.625 and 230Th/238U and 234U/238U at equilibrium, except age marked with

an asterisk, which is corrected by using measured values on insoluble residue230Th/232Th = 0.8561 T 0.0039, and age marked with a dagger, which iscorrected by using measured values on insoluble residue 230Th/232Th =0.9390 T 0.0077.

Sample

BIG-UTh-Site Description 230Th/238U 234U/238U 230Th/232Th

Uncorrected

age (ky)

Corrected

age (ky)

Minimum ages

O-53 Altamira

Overlays red spotted outline

horse of Techo de los

Polícromos chamber

0.2884 T 0.0013 1.5471 T 0.0026 107.07 T 0.20 22.26 T 0.11 22.11 T 0.13

O-80 El Castillo

Overlays black outline drawing

of indeterminate animal of

corridor of Techo de las Manos

0.7879 T 0.0047 3.9828 T 0.0073 30.01 T 0.15 23.43 T 0.16 22.88 T 0.27

O-58 El CastilloOverlays red stippled negative

hand stencil of Techo de las Manos0.5272 T 0.0020 2.5774 T 0.0049 222.70 T 0.49 24.42 T 0.11 24.34 T 0.12

O-21 Tito Bustillo

Red pigment associated with

anthropomorphic figure of

Galería de los Antropomorfos

0.6252 T 0.0031 1.8038 T 0.0037 2.17 T 0.01 44.94 T 0.2930.8 T 5.6

29.65 T 0.55*

O-69 El CastilloLarge red disk of

Galería de los Discos0.7512 T 0.0029 2.7072 T 0.0051 788.2 T 5.5 34.28 T 0.17 34.25 T 0.17

O-50 AltamiraLarge red claviform-like

symbol of Techo de los Polícromos0.4933 T 0.0024 1.6594 T 0.0030 17.473 T 0.068 37.60 T 0.23 36.16 T 0.61

O-82 El Castillo

Sample overlays red negative

hand stencil, and underlies yellow

outline bison of Panel de las Manos

0.5112 T 0.0029 1.6970 T 0.0035 48.81 T 0.49 38.15 T 0.27 37.63 T 0.34

O-83 El CastilloOverlays large red stippled disk

of Panel de las Manos0.3573 T 0.0022 1.1048 T 0.0020 28.64 T 0.29 42.38 T 0.33 41.40 T 0.57

Maximum ages

O-87 El Castillo

Underlies large red disk of

Galería de los Discos

(same panel as O-69)

0.7969 T 0.0038 2.7432 T 0.0051 61.24 T 0.61 36.11 T 0.21 35.72 T 0.26

O-48 Tito Bustillo

Underlies red anthropomorph

figure of Galería de los

Antropomorfos (see also O-21)

0.5281 T 0.0038 1.6895 T 0.0042 7.260 T 0.047 39.85 T 0.3636.2 T 1.5

35.54 T 0.39†

Fig. 2. U-series ages representing minimum ages for the cave art that we have sampled. Also shown isthe pre-Magdalanian radiocarbon chronology from excavated cultural horizons in northern Spain (witharbitrary y-axis offsets for clarity).

15 JUNE 2012 VOL 336 SCIENCE www.sciencemag.org1410

RESEARCH ARTICLE

on J

une 1

4,

2012

ww

w.s

cie

ncem

ag.o

rgD

ow

nlo

aded fro

m

famous Techo de los Polícromos (Fig. 4); thus,

it must be considerably older and at least Au-

rignacian in age. The previous chronological

schemes all placed the panel’s polychrome im-

ages later, in the Magdalenian (5). Our date thus

indicates earlier use of the cave and support, and

may extend, the long chronology. Although a

few small claviforms on the same panel are su-

perimposed on the polychromes and thus were

painted afterward, our result for the larger version

of the sign confirms Breuil’s suggestion that red

claviforms (or claviform-like symbols) originated

before the Solutrean and probably in the Aurigna-

cian and had a long chronological currency (24).

El Castillo. The caves of El Castillo, along

the Pas river in northern Spain, also contain more

than 100 images in multiple chambers. We ob-

tained an age of >22.6 ky (O-80) for the black

outline drawing of an indeterminate animal, which

demonstrates that it was painted at least during

the Solutrean. Traditional chronological schemes

attribute most of the black animal figures to Mag-

dalenian times (25) as a succession to the earlier

red figures. This result suggests an earlier chro-

nology for at least some of the black figures.

We also dated calcite overlying one red disk

in the Corredor de los Puntos (fig. S7) and un-

derneath another one, providing a minimum age

of 34.1 ky (O-69) and a maximum age of 36.0 ky

(O-87). Given the homogeneity in technique and

location of the various large disks, it seems rea-

sonable to assume they represent a single episode

of painting. If so, the dates constrain the paintings

to the latest part of the Aurignacian.

Hand stencils (O-58 and -82) are found in

numerous caves in France and Spain. They are

thought to have been made by blowing pigment

onto the hand placed against the cave wall. They

are usually assumed to be mainly or exclusively

Mid Upper Paleolithic (Gravettian) in age, and,

where available, direct radiocarbonmeasurements

support this notion (ages range from ~29,000

to 22,000 14C yr B.P. for paintings at Gargas,

Cosquer, Labattut, Fuente de Salin, and Paglicci)

(26–28).We dated two hand stencils at El Castillo.

One date of >24.2 ka for O-58, close to the

Solutrean/Gravettian boundary, is consistent with

the stylistic attribution of hand stencils to the

Gravettian. But the other date from a hand stencil

from the Panel de las Manos (O-82) is older,

Fig. 3. A time line of thecave art dated. A single ar-row represents a minimumage, but, where two datesare indicated, bothmaximumandminimumageshavebeenobtained. The error bars forO-21 reflect the variation re-sulting from the two differentmethods of detrital correc-tion (11). Larger versions ofthese images showing sam-ple locations are availablein the supplementary mate-rials, figs. S2 to S12.

Fig. 4. The Techo de los Polícromos, Altamira Cave, showing the red spotted outline horse (sample O-53)and the larger red claviform-like symbol (sample O-50). [Tracing credit: National Museum and ResearchCentre of Altamira/Pedro Saura]

www.sciencemag.org SCIENCE VOL 336 15 JUNE 2012 1411

RESEARCH ARTICLE

on J

une 1

4,

2012

ww

w.s

cie

ncem

ag.o

rgD

ow

nlo

aded fro

m

>37.3 ka, and from the Aurignacian period or

earlier. This date considerably increases the antiq-

uity of this motif and implies that depictions of

the human hand were among the oldest art known

from Europe. One of several large red disks

(O-83) nearby, also made by using a blowing

technique, has a minimum age of 40.8 ky, sup-

porting this interpretation. Because the disk and

hand stencil are in the same panel (Fig. 5), it is

possible that the older age of 40.8 ky is the min-

imum age of the entire composition, including

rectangular and oval signs, about 40 hand sten-

cils, and dozens of large disks, indicating intense

artistic activity in pre-Gravettian times.

Tito Bustillo. The Galeria de los Antropo-

morfos in the cave of Tito Bustillo (Ribadesella)

contains a scarf stalactite, about 3 m above the

present ground surface, painted on two sides with

two anthropomorphic figures and red pigment

along the vertices of the stalactite. We dated cal-

cite on top of this red pigment, providing mini-

mum ages with two different methods of detrital

correction (11) of 29.6 and 25.2 ky, and a sample

drilled into a recent break, providing maximum

ages of 35.5 and 37.7 ky, (samples O-21 and -48).

These results put the possible age of the painting

from the Gravettian (either early or late) to the

Late Aurignacian. Excavation in the Galería de

los Antropomorfos yielded charcoal radiocarbon

dated to 32,990 T 450 14C yr B.P. (38,729 to

36,665 cal yr B.P.) (29) associated with a short-

lived and nondomestic human occupation as far

back as the Aurignacian, represented by pigment,

crushed bones, and deliberately selected flat stones.

Our result confirms that this cave hosts a long

artistic tradition. These results may imply an old

age for early paintings that are partly covered in

the Panel de los Polícromos, elsewhere in the

cave, which is thought to have been painted in the

Magdalenian (7, 29).

Discussion

Our dates show that artistic activity began at least

in the Aurignacian period, in two distinct caves,

and in one case at least as far back as 40.8 ka.

The large red disk from El Castillo (O-83)

currently represents the earliest dated example of

European cave art. Its age is ~4000 years earlier

than the claimed age of the art at Grotte Chauvet,

France. The oldest radiocarbon date there is

32,410 T 720 14C yr BP (35,300 to 38,827 cal

yr B.P.), although others cluster around 35,000

cal yr B.P. (30, 31).

Pre-Gravettian cave painting should be evi-

dent elsewhere in Europe. Blocks presumed to be

a former cave wall, with red pigment (one in-

terpreted as an anthropomorphic figure), have

been found in Aurignacian layers in Fumane

Cave, Italy (32), and in several rockshelters of the

Vezere River, Dordogne (33). The pre-Gravettian

signs (large dots, disks, and the claviform-like

symbol) and hand stencils that we dated are sim-

ilar to many found in other caves across Europe,

although use of particular symbols may differ

regionally (34). The apparent lack of further pre-

Gravettian examples elsewhere in Europe ismore

likely the result of targeting charcoal-based black

pigments with radiocarbon dating and the result-

ing ambiguities than a genuine absence of pre-

Gravettian images.

Our results are consistent with the notion that

there was a gradual increase in technological and

graphic complexity over time, as well as a grad-

ual increase in figurative images. Our earliest

dates (pre-Gravettian) are for art that is nonfigu-

rative and monochrome (red), supporting the no-

tion that the earliest expression of art in Western

Europe was less concerned with animal depic-

tions and characterized by red dots, disks, lines,

and hand stencils. If the earliest cave paintings

appeared in the region shortly before 40.8 ka,

this would, assuming that the Proto-Aurignacian

cultural complexwasmade exclusively byHomo

sapiens, support the notion that cave art co-

incided with their arrival in western Europe

~41.5 ka and that the exploration and deco-

rating of caves was part of their cultural package.

However, because the 40.8-ky date for the disk

is a minimum age, it cannot be ruled out that

the earliest paintings were symbolic expres-

sions of the Neandertals, which were present

in Cantabrian Spain until at least 42 ka (13, 15).

References and Notes1. A. Sinclair, Nature 426, 774 (2003).2. J. Zilhão, J. Archaeol. Res. 15, 1 (2007).3. N. J. Conard, Proc. Natl. Acad. Sci. U.S.A. 107, 7621 (2010).4. J. Fortea, in Pitture Paleolitiche nelle Prealpi venete.

Grotta di Fumane e Riparo Dalmeri, A. Broglio,G. Dalmeri, Eds. (Museo Civico di Storia Naturaledi Verona, Verona, 2005), pp. 89–99.

5. A. Leroi-Gourhan, B. Delluc, G. Delluc, Préhistoire del’art occidental (Mazenod, Paris, 1995).

6. P. B. Pettitt, A. W. G. Pike, J. Archaeol. Method Theory14, 27 (2007).

Fig. 5. The Panel de las Manos, El Castillo Cave, showing the location of samples O-82 overlaying anegative hand stencil, and O-83 overlaying a large red stippled disk. The tracing is taken from (35).

15 JUNE 2012 VOL 336 SCIENCE www.sciencemag.org1412

RESEARCH ARTICLE

on J

une 1

4,

2012

ww

w.s

cie

ncem

ag.o

rgD

ow

nlo

aded fro

m

7. F. J. Fortea Pérez, Bull. Soc. Préhistorique AriègePyrénées LVII, 7 (2002).

8. A. W. G. Pike et al., J. Archaeol. Sci. 32, 1649(2005).

9. D. L. Hoffmann, Int. J. Mass Spectrom. 275, 75(2008).

10. D. L. Hoffmann et al., Int. J. Mass Spectrom. 264,97 (2007).

11. Materials and methods are available as supplementarymaterials on Science Online.

12. P. J. Reimer et al., Radiocarbon 51, 1111 (2009).13. J. Zilhão, Pyrenae 37, 7 (2006).14. J. Zilhão et al., PLoS ONE 5, e8880 (2010).15. J. Maroto et al., Quat. Int. 247, 15 (2012).16. L. G. Straus, Evol. Anthropol. 14, 145 (2005).17. J. Zilhão, T. Aubry, F. Almeida, in Les faciès leptolithiques

du nord-ouest méditerranéen: milieux naturels et culturels.

XXIVe Congrès Préhistorique de France. Carcassonne,

26-30 Septembre 1994, D. Sacchi, Ed. (SociétéPréhistorique Française, Paris, 1999), pp. 165–183.

18. C. Renard, World Archaeol. 43, 726 (2011).19. I. J. Fairchild, S. Frisia, A. Borsato, A. F. Tooth, in

Geochemical Sediments and Landscapes, D. J. Nash,S. J. McLaren, Eds. (Blackwell, Oxford, 2007), pp. 200–245.

20. A. Beltran, The Cave of Altamita (Harry Abrams,New York, 1999)

21. C. González-Sainz, R. Cacho Toca, R. Fukazawa, ArtePaleolítico en la Región Cantábrica (Universidad deCantabria, Cantabria, Spain, 2003).

22. F. Bernaldo de Quirós, Proc. Prehistoric Soc. 57, 81(1991).

23. C. Heras Martín, R. Montes Barquín, J. A. Lasheras,P. Rasines, P. Fatás Monforte, Cuadernos de Arte Rupestrede Moratalla 4, 117 (2008).

24. H. Breuil, E. Cartailhac, H. Obermaier, M. E. Boyle,The Cave of Altamira at Santillana del Mar, Spain

(Tip. de Archivos, Madrid, 1935).25. H. Breuil, Quatre cents siècles d’art pariétal. Les cavernes

ornées de l’age du renne (Centre d’Etudes deDocumentation Préhistoriques, Montignac, Paris, 1952).

26. M. Lorblanchet, Les Grottes Ornées de la Préhistoire.Nouveaux Regards (Editions Errance, Paris, 1995).

27. F. Djindjian, in Art Mobilier Paléolithique Supérieur enEurope Occidental, A-C. Welté, E. Ladier, Eds. [Actes duColloque 8.3, Congrès de l’UISPP (Union Internationaledes Sciences Préhistoriques et Protohistoriques), Universitéde Liège, Liège, Belgium, 2004], pp. 249–256.

28. J. Clottes, J. Courtin, The Cave Beneath the Sea: PalaeolithicImages at Cosquer (H. N. Adams, New York, 2006).

29. R. de Balbín, J. J. Alcolea, M. A. González, in El ArtePrehistórico Desde Los Inicios del s. XXI. Primer Symposium

Internacional de Arte Prehistórico de Ribadesella,R. de Balbín, P. Bueno, Eds. (Asociación Cultural deAmigos de Ribadesella, Ribadesella, Spain, 2003).

30. J. Clottes, Return to Chauvet Cave: Excavating the

Birthplace of Art (Thames & Hudson, London, 2003).31. P. B. Pettitt, P. Bahn, C. Züchner, in An Enquiring Mind:

Studies in Honor of Alexander Marshack, P. Bahn, Ed.

(American School of Prehistoric Research MonographSeries, Oxbow Books, Oxford, UK, 2009), pp. 239–262.

32. A. Broglio, M. de Stefani, F. Gurioli, M. Peresani,Int. Newsl. Rock Art 44, 1 (2006).

33. B. Delluc, G. Delluc, L’Art Pariétal Archaïque en Aquitaine(Gallia Préhistoire XXVIII supplément, CNRS, Paris, 1991).

34. M. García-Diez, J. A. Mujika Alustiza, M. Sasieta,J. Arruabarrena, J. Alberdi, Int. Newsl. Rock Art 60, 13(2011).

Acknowledgments: This research was funded by a grantto A.W.G.P. from the Natural Environmental ResearchCouncil (NE/F000510/1). C. Taylor performed the samplepreparation and assisted in collecting samples in the fieldalong with C. Hinds, S. White, and S. Paine. We thank thegovernments of Asturias and Cantabria for grantingpermission to sample the cave art and B. Hanson foreditorial guidance. The data described are presented inthe supplementary materials.

Supplementary Materialswww.sciencemag.org/cgi/content/full/336/6087/1409/DC1Materials and MethodsFigs. S1 to S12Table S1References (35–39)

1 February 2012; accepted 25 April 201210.1126/science.1219957

REPORTS

Spin-Orbit EchoN. Sugimoto1,2 and N. Nagaosa1,2*

Preserving and controlling the quantum information content of spins is a central challenge ofspintronics. In solids, the relativistic spin-orbit interaction (SOI) leads to a finite spin lifetime.Here, we show that spin information is preserved by the hidden conserved “twisted spin” andsurvives elastic disorder scatterings. This twisted spin is an adiabatic invariant with respect to aslow change in the SOI. We predict an echo phenomenon, spin-orbit echo, which indicatesthe recovery of the spin moment when the SOI is tuned off adiabatically, even after spinrelaxation has occurred; this is confirmed by numerical simulations. A concrete experiment intwo-dimensional semiconductor quantum wells with Rashba-Dresselhaus SOI is proposed toverify our prediction.

For the manipulation of spin and spin

current by electric fields in spintronics,

the relativistic spin-orbit interaction (SOI)

is essential because it connects the orbital motion

and the spin of an electron (1). This interaction,

however, destroys the rotational symmetry in the

spin space and the consequent spin conservation

law. In the presence of the SOI, there are several

mechanisms to relax the spins by changing their

direction at and between impurity scatterings (Fig.

1A) (2–8); spins decay with a (phenomenological)

spin lifetime ts, which is typically on the order of

(or less than) 1 ns [(Fig. 1B, (i) to (iv)]. This is a

serious issue for applications because the transfer

and storage of information for a long time/distance

is a crucial requirement for device function.

It has long been recognized that the SOI is

closely related to the parallel transport and rotat-

ing frame comoving with the electron in the con-

text of Thomas precession (9). This naturally

leads to the formulation of the SOI in terms of the

SU(2) spin gauge field %Am (m = t, x, y, z) con-

necting the neighboring frames in the non-

relativistic approximations to the Dirac equation

(10–12). Based on this idea, there have been

many attempts to find a conserved quantity by

generalizing the spin (11–17). It has been known

(11, 12, 18) that the electronic spin is only covar-

iantly conserved as D0sa +D·ja = 0, a spin density

sa, and a spin current density ja (a = x, y, z); i.e.,

the continuity equation where the usual deriva-

tive ∂m is replaced by the covariant derivative

Dm :¼ (D0, D) holds. This means that the conser-

vation law is satisfied in the comoving spin frame

with the electron’smotion but not in the laboratory

frame. In a notable recent advance related to this

concept, experiments were performed in

semiconductor quantum wells with Rashba

(a) and Dresselhaus (b) SOIs described by

HRD ¼ p2

2mþ aðpysx− pxsyÞ þ bðpxsx − pysyÞ,

where HRD is the Rashba-Dresselhaus Hamil-

tonian, p is the momentum, m is the mass of an

electron, ss are the Pauli matrices, and ħ = 1. In

this system, the lifetime of a Fourier compo-

nent of the spin Sq for q ¼ ð2ffiffiffi

2p

ma; 2ffiffiffi

2p

maÞ[persistent spin helix (PSH)] is observed to be

enhanced when the condition a = b is satisfied

(19, 20). In this case, the SU(2) gauge field

strength %Fmnºða2 − b2Þsz is zero; that is, the

SOI vanishes in the rotated spin frame.

We explicitly construct a conserved twisted

spin in a generic situation when the electric field

is regarded as a background static field (Fig. 1A).

Furthermore, we study the adiabatic invariance of

this twisted spin and predict a phenomenon

called the spin-orbit echo [Fig. 1B, (iv) to (vii)].

The twisted spin preserves the information on the

spin, and as the SOI is switched off adiabatically,

the spin is recovered to coincide with the con-

served twisted spin.

We consider a noninteracting electron system

whose Hamiltonian is given by (10–12)

H ¼ 1

2mp −

e

mc2%A

!2

þ V ðxÞ ð1Þ

where %A ¼ ∑i,aAai eisa=2, where Aa

i :¼ ∑leialEl

describes the SOI, ei is the unit vector, and V(x) is

a scalar potential that includes the periodic crystal

potential and the disorder potential. Here, –e repre-

sents the charge of an electron; c is the velocity of

1Department of Applied Physics, University of Tokyo, Tokyo113-8656, Japan. 2Cross-Correlated Materials Research Groupand Correlated Electron Research Group, RIKEN, Saitama 351-0198, Japan.

*To whom correspondence should be addressed. E-mail:nagaosa@ap.t.u-tokyo.ac.jp

www.sciencemag.org SCIENCE VOL 336 15 JUNE 2012 1413

on J

une 1

4,

2012

ww

w.s

cie

ncem

ag.o

rgD

ow

nlo

aded fro

m

www.sciencemag.org/cgi/content/full/336/6087/1409/DC1

Supplementary Materials for

U-Series Dating of Paleolithic Art in 11 Caves in Spain

A. W. G. Pike,* D. L. Hoffmann, M. García-Diez, P. B. Pettitt, J. Alcolea, R. De Balbín, C. González-

Sainz, C. de las Heras, J. A. Lasheras, R. Montes, J. Zilhão

*To whom correspondence should be addressed. E-mail: alistair.pike@bristol.ac.uk

Published 15 June 2012, Science 336, 1409 (2012)

DOI: 10.1126/science.1219957

This PDF file includes:

Materials and Methods

Figs. S1 to S12

Table S1

References

Supplementary Materials:

Materials and Methods

The Uranium-series disequilibrium method

The U-series disequilibrium method is based on the radioactive decay of radionuclides within the

naturally occurring decay chains. There are three such decay chains, each starts with an actinide

nuclide (i.e., 238

U, 235

U, and 232

Th) having a long half live (all have T1/2 >7x108 years) and ultimately

ends with different stable isotopes of lead. For dating speleothems, we make use of an initial

elemental fractionation between Th and U in the 238

U decay series when carbonate bedrock is

dissolved. Differential solubility between uranium and its long lived daughter isotope 230

Th means

that calcite precipitates (e.g. stalagmites, stalactites and flowstones) contain traces of uranium but, in

theory, no 230

Th. Over time, there is ingrowth of 230

Th from the radioactive decay of 238

U until

radioactive equilibrium is reached where all isotopes in the series are decaying at the same rate. It is

the degree of disequilibrium (measured as 230

Th/238

U activity ratio) that can be used together with the

activity ratio of the two U isotopes 234

U/238

U to calculate the age of the calcite precipitation. Natural

processes usually also cause a disequilibrium between 238

U and 234

U, so the age since formation of a

calcite sample is calculated iteratively from measurements of 234

U/238

U and 230

Th/238

U (36).

An additional problem is the incorporation of detritus in the precipitating calcite. This can be from

wind-blown or waterborne sediments. Detrital sediments will bring U and Th and usually will result

in the apparent age of a contaminated sample to be an overestimate of the true age. However, the

presence of the common thorium isotope, 232

Th, indicates the presence of contamination, and there are

several methods to correct the U-series date for it. An indication of the degree of detrital

contamination is expressed as 230

Th/232

Th activity, with high values (>20) indicating little or no effect

on the calculated date and low values (<20) indicating the correction on the date will be significant.

For very low values of 230

Th/232

Th (i.e <5), the calculated age will be dominated by the assumptions

used to correct for the detritus, so we employ two different correction strategies. For samples with

minor and moderate levels (230

Th/232

Th > 5) of contamination, we correct using an assumed detrital

activity ratio of 232

Th/238

U=1.250±0.625, typical of upper crustal silicates (37) and assume 230

Th and

U isotopes are in equilibrium (i.e. 230

Th/238

U=1.0; 234

U/238

U=1.0). Note the conservative error on this

assumption. For samples with high levels of detritus, we attempt to measure typical detrital 230

Th/232

U

on the insoluble residues from the calcite samples and report both a date using a crustal silicate

correction and a date using the measured detrital value (see for example samples O-21 and O-48).

While the date obtained using measured detritus values agrees within error in both cases with the date

using an average crustal silicate, we must be cautious in using dates corrected using the insoluble

detritus. Soluble Th in the detritus, or the adsobtion of authigenic Th onto the detritus may introduce

uncertainty in determining the true detrital 230

Th/232

Th (38). This uncertainty is unlikely to be as large

as the 50% uncertainty we are using for our assumed detrital value, but it is likely to be greater than

the uncertainty we give. To be cautious therefore, we base our interpretation of the dates for samples

O-21 and O-48 on dates corrected with our assumed rather than measured detrital value.

Method

In order to be certain that a minimum or maximum age is obtained, it is essential to select calcite

deposits that have an unambiguous stratigraphic relationship with the painting or engraving. Each

potential location was inspected with a hand lens and locations were sampled only where the painting

was clearly covered with calcite, or where the underlying calcite was accessible. The calcite deposits

ranged from small discontinuous nodules (e.g. Fig. S6), ‘trickles’ of flowstone (e.g. Fig. S3), through

to the occasional continuous flowstone crust. Powdery and soft deposits were avoided. The majority

of samples were removed by scraping with a scalpel, catching the scrapings in a cleaned plastic tray.

The calcite was removed in spits, creating aliquots of sample. This allowed regular inspection of the

scraped surface and the aliquots of calcite in order to (a) avoid unintended inclusion of scraped

pigment, which would contaminate the sample and (b) make sure the sample removed was still

entirely from above the painting, so as not to damage it. Aliquots contaminated with pigment or

visible detritus were discarded and the remaining aliquots combined to give sample masses of 10-

100mg. Where a sufficient thickness of calcite was present (>2mm), 2 samples were removed,

representing the upper and lower portions of the calcite crust. In all cases, the dates of these fall in the

correct stratigraphic order, demonstrating the integrity of the calcite (Fig. S1). In some cases where

the formation was stalactitic, samples were cut with a diamond cutting wheel, or drilled with a carbide

drill bit. A further demonstration of the reliability of the technique comes from the distribution of

results, which show that the formation ages for calcite on top of art fall between a few hundred years

and 40.8 ka (Fig. 2). Since calcite formation has been ongoing in most caves over a period beyond the

limit of the U-Th method (c. 500 ka), this distribution would not be expected if the stratigraphic

relationship between the art and the calcite was insecure.

Samples were initially inspected under a low power microscope and, where possible, any obvious

particles of detritus were removed. The sample was weighed in a Teflon beaker. A few drops of

milliQ 18MΩ water were added, and the sample was dissolved by further stepwise addition of 7N

HNO3. A mixed

229Th/

236U spike was added and left for a few hours to equilibrate. Where appropriate,

any insoluble residue was removed by centrifuge. The sample solution was dried by placing the

beaker on a hotplate. When nearly dry the sample was treated with 100μl 6N HCl and 55μl H2O2 and

left until dry. Finally, the sample was re-dissolved in 600μl 6N HCl ready for the ion exchange

columns.

U and Th were separated from the sample matrix using ion exchange chromatography and a two

column procedure (9). The first column separates U from Th and the second purifies the two fractions.

We use 600μl of pre-cleaned Bio Rad AG1x8 resin. The sample is introduced into the first column in

6N HCl. The Th fraction is collected immediately as it passes directly through the column. U is then

eluted using 1N HBr followed by 18MΩ water. After drying down the two fractions were redissolved

in 7N HNO3 and separately passed down the column for purification. Th is eluted with 6N HCl and U

is eluted with 1N HBr. The elutants were dried then redissolved in 0.6N HCl ready for analysis.

U and Th isotope measurements were undertaken using a ThermoFinnigan Neptune Multi-Collector

(MC) Inductively Coupled Plasma Mass Spectrometer (ICPMS). Instrumental biases are assessed and

corrected by adopting a standard - sample bracketing procedure to derive correction factors e.g. for

mass fractionation effects. U and Th solutions are measured separately; NBL-112a is used for U

isotope measurements as the bracketing U-standard and an in-house 229

Th-230

Th-232

Th standard

solution for Th measurements. Further details of our MC-ICPMS procedures can be found in

references (9, 10). U-series dating of speleothems is described in more detail in reference (39).

Minimum ages are quoted as measured age minus 2σ and maxiumum ages as measured age plus 2σ.

Date Reporting Conventions

Unlike radiocarbon dates, U-series disequilibrium produces results in calendar years. To distinguish

between radiocarbon years and U-series results we quote U-series ages as ky (thousands of years),

uncalibrated radiocarbon dates as 14C yr BP (radiocarbon years before present, the present being the

year 1950 AD), and calibrated radiocarbon dates as cal yr BP (calibrated years before present,

equivalent to calendar years). For dates (i.e. points in time in the past) we use ka (thousands of years

before today).

Supplementary Figures

Fig. S1. Uranium series dates on paired samples. Aliquots of samples were removed, representing the

upper and lower portions of the calcite crust and dated separately. In all cases, the dates of the upper

portions are younger than the lower portions (i.e. following stratigraphic deposition of the calcite),

demonstrating the integrity of the samples.

Fig. S2. Sample O-53, overlays red ‘spotted outline’ horse ofTecho de los Polícromos chamber,

Altamira Cave. The location of this symbol on the Techo de los Polícromos is shown in Fig. S9.

Image ©National Museum and Research Centre of Altamira / Pedro Saura.

Fig. S3. Sample O-80, El Castillo Cave, overlays black outline drawing of an indeterminate animal in

corridor of Techo de las Manos.

Fig. S4. Sample O-58 overlays red stippled negative hand stencil of Techo de las Manos, El Castillo

Cave. Note red pigment revealed under sample.

Fig. S5. Anthropomorph figure of Galería de los Antropomorfo, Tito Bustillo Cave. Sample O-58

overlays red pigment of vertex of scarf stalactite; sample O-48 is drilled from a recent break in the

stalactite providing a maximum age for the figure.

Fig. S6. Sample O-69 overlays large red disk of Galería de los Discos, El Castillo Cave.

Fig. S7. Galería de los Discos, El Castillo Cave. Sample O-87 underlies a large red disk, and provides

a maximum age. (Image © Consejería de Cultura, Turismo y Deporte, Gobierno de Cantabria / Pedro

Saura)

Fig. S8. Sample O-50 overlays large red claviform-like symbol on the Techo de los Polícromos,

Altamira Cave. The location of this symbol on the Techo de los Polícromos is shown in Fig. S9.

Image © The National Museum and Research Centre of Altamira / Pedro Saura.

Fig. S9. Digital reconstruction (top) and drawing (bottom) of the Techo de los Polícromos, Altamira

Cave, showing the location of the claviform-like symbol (sample O-50) and the red spotted outline

horse (sample O-53). Images ©The National Museum and Research Centre of Altamira / Pedro Saura.

O-53

O-50

Fig. S10. Sample O-82 overlays red negative hand stencil, and underlies yellow outline bison of

Panel de las Manos, El Castillo Cave. See also Fig. S12. Note the red pigment revealed under the

sample.

Fig. S11. Sample O-83 overlays large red stippled disk on the Panel de las Manos, El Castillo Cave. The age of

>40.8 ky makes this the oldest known cave art in Europe. The pre-Gravettian date for a hand stencil on the same

panel (O-82 at >37.3ky) and the similarity in painting technique may indicate that all the stippled paintings on

this panel are contemporary representing more than 50 motifs (see Fig. S12). The yellow bison is superimposed

on this composition and represents a later addition to the panel.

Fig. S12. The Panel de las Manos, El Castillo Cave showing the location of samples O-82 overlaying

a negative hand stencil, >37.3 ky , and O-83 overlaying a large red stippled disk, >40.8 ky . The

tracing in the lower panel is taken from ref (35).

O-82

O-83

Sa

mp

le

BIG

-UT

h-

Sit

e

Des

crip

tio

n

230T

h/2

38U

234U

/238U

230T

h/2

32T

h

Un

corr

ecte

d

Ag

e (k

y)

Co

rrec

ted

Ag

e

(ky

)

O-3

0

Tit

o B

ust

illo

O

ver

lies

red

ho

rse,

En

sem

ble

X

0.0

015

21 ±

0.0

00

042

0.8

791

± 0

.00

16

9

.60

± 0

.29

0

.18

91

± 0

.00

53

0

.17

34

± 0

.00

95

O-1

01

L

a P

asie

ga

Over

lies

red

bo

vid

,

Pas

iega

C

0.0

196

2 ±

0.0

004

7

2.1

559

± 0

.00

43

2

.95

9 ±

0.0

75

0

.99

8 ±

0.0

24

0

.73

± 0

.14

O-1

03

L

a P

asie

ga

Over

lies

red

meg

alo

cero

s,

Pas

iega

B

0.0

486

6 ±

0.0

004

3

3.0

720

± 0

.00

58

4

0.1

7 ±

0.4

7

1.7

41 ±

0.0

16

1

.70

6 ±

0.0

23

O-1

09

L

a P

asie

ga

Over

lies

red

und

eter

min

ed

figure

, P

asie

ga

B

0.1

314

6 ±

0.0

006

6

6.3

84 ±

0.0

10

2

25

.5 ±

1.9

2

.26

6 ±

0.0

12

2

.25

8 ±

0.0

13

O-8

8

El

Cas

till

o

Over

lies

sm

all

red

do

t, G

ran

Sa

la

0.0

865

3 ±

0.0

004

9

4.0

458

± 0

.00

72

1

16

.03

± 0

.81

2

.35

5 ±

0.0

14

2

.33

9 ±

0.0

16

O-1

06

L

a P

asie

ga

Over

lies

red

und

eter

min

ed

figure

, P

asie

ga

B

0.1

425

0 ±

0.0

009

0

6.2

13 ±

0.0

11

6

95

.5 ±

5.9

2

.52

6 ±

0.0

17

2

.52

3 ±

0.0

17

O-7

1

Alt

amir

a

Over

lies

bla

ck i

bex

,

La

Ho

ya

0.0

532

0 ±

0.0

008

1

1.6

567

± 0

.00

30

3

.96

4 ±

0.0

58

3

.55

7 ±

0.0

55

2

.85

± 0

.35

O-1

07

L

a P

asie

ga

Over

lies

red

bis

on,

Pas

iega

B

0.1

557

± 0

.00

11

5

.03

41

± 0

.00

80

2

4.4

2 ±

0.2

7

3.4

17 ±

0.0

25

3

.30

7 ±

0.0

55

O-1

08

L

a P

asie

ga

Over

lies

red

bis

on,

Pas

iega

B

0.1

308

± 0

.00

10

4

.29

12

± 0

.00

75

10

0.5

8 ±

0.9

6

3.3

68 ±

0.0

27

3

.34

2 ±

0.0

29

O-1

05

L

a P

asie

ga

Over

lies

red

ho

rses

,

Pas

iega

B

0.1

168

2 ±

0.0

006

5

3.2

383

± 0

.00

55

1

06

.09

± 0

.96

3

.99

6 ±

0.0

24

3

.96

7 ±

0.0

27

O-1

10

L

a P

asie

ga

Over

lies

red

ho

rse,

Pas

iega

B

0.3

140

± 0

.00

19

7

.85

7 ±

0.0

14

3

8.9

1 ±

0.3

4

4.4

29 ±

0.0

29

4

.34

0 ±

0.0

44

O-7

3

La

Pas

iega

Over

lies

red

tria

ng

ula

r sy

mb

ol,

Pas

iega

C

0.1

399

7 ±

0.0

006

2

3.3

805

± 0

.00

54

3

08

.1 ±

1.8

4

.59

6 ±

0.0

22

4

.58

5 ±

0.0

22

O-1

02

L

a P

asie

ga

Over

lies

bla

ck i

bex

,

Pas

iega

C

0.1

058

± 0

.00

11

2

.17

15

± 0

.00

40

3

8.5

9 ±

0.4

5

5.4

33 ±

0.0

59

5

.32

3 ±

0.0

78

O-7

6

La

Pas

iega

Over

lies

red

clav

ifo

rm,

Pas

iega

B

0.2

330

± 0

.00

45

4

.58

23

± 0

.01

9

92

.6 ±

1.9

5

.66

± 0

.11

5

.61

5 ±

0.1

16

O-4

6

Alt

amir

a

Over

lies

red

tech

tifo

rm,

sect

or

III

0.0

798

0 ±

0.0

004

7

1.4

959

± 0

.00

26

4

0.2

9 ±

0.3

5

5.9

69 ±

0.0

38

5

.85

4 ±

0.0

68

O-8

4

El

Cas

till

o

Over

lies

red

dee

r,

Ga

lerí

a d

el B

iso

nte

’ 0

.10

68

± 0

.00

22

1

.47

07

± 0

.00

27

2

.81

2 ±

0.0

61

8

.20

± 0

.18

5

.9 ±

1.1

O-7

7

Co

val

anas

O

ver

lies

red

bo

vid

0

.10

46

± 0

.00

11

1

.85

66

± 0

.00

37

7

9.4

9 ±

0.8

6

6.3

04 ±

0.0

67

6

.24

2 ±

0.0

73

O-7

8

San

tián

O

ver

lies

red

“hand

-

like”

sy

mb

ol

0.0

767

6 ±

0.0

005

1

1.3

083

± 0

.00

27

1

40

.4 ±

1.1

6

.58

8 ±

0.0

47

6

.55

2 ±

0.0

50

O-2

2

Tit

o B

ust

illo

Red

pig

ment

asso

ciat

ed w

ith

anth

rop

om

orp

hic

figure

, G

ale

ría

de

los

An

tro

po

mo

rfo

s

0.1

872

± 0

.00

27

1

.97

57

± 0

.00

40

2.1

15 ±

0.0

31

1

0.7

9 ±

0.1

7

6.9

± 1

.9

O-9

8

La

Pas

iega

Over

lies

sm

all

red

do

t, P

asie

ga

C

0.2

337

± 0

.00

12

3

.66

45

± 0

.00

65

1

55

.8 ±

1.3

7

.14

2 ±

0.0

41

7

.10

7 ±

0.0

44

O-6

8

El

Cas

till

o

Over

lies

bla

ck

ho

rse,

El

Pa

so

0.3

543

± 0

.00

26

4

.99

39

± 0

.00

90

1

5.9

2 ±

0.1

0

7.9

64 ±

0.0

63

7

.58

± 0

.16

O-5

6

Co

val

anas

O

ver

lies

red

dee

r 0

.13

61

2 ±

0.0

009

9

1.8

976

± 0

.00

34

2

6.2

6 ±

0.2

4

8.0

83 ±

0.0

62

7

.85

± 0

.13

O-6

0

San

tián

Over

lies

red

co

lor

conce

ntr

atio

n o

n

stal

agm

itic

pil

lar,

Mai

n C

orr

ido

r

0.1

490

± 0

.00

29

1

.90

56

± 0

.00

77

9

.10

± 0

.20

8.8

4 ±

0.1

8

8.0

9 ±

0.3

9

O-9

1

El

Cas

till

o

Over

lies

bla

ck

bo

vid

, G

ale

ría

del

Bis

on

te

0.3

136

± 0

.00

28

4

.18

93

± 0

.00

88

5

7.6

6 ±

0.5

8

8.4

18 ±

0.0

80

8

.30

6 ±

0.0

91

O-7

4

La

Pas

iega

Over

lies

yel

low

do

ub

le a

rch m

oti

f,

Pas

iega

C

0.3

010

± 0

.00

13

3

.80

29

± 0

.00

64

3

6.6

2 ±

0.1

2

8.9

18 ±

0.0

41

8

.73

2 ±

0.0

85

O-1

00

L

a P

asie

ga

Over

lies

red

dee

r,

Pas

iega

C

0.2

881

± 0

.00

19

3

.62

86

± 0

.00

78

2

81

.5 ±

2.3

8

.94

8 ±

0.0

63

8

.92

4 ±

0.0

63

O-8

9

El

Cas

till

o

Over

lies

red

‘b

ell’

,

Pa

nel

de

los

Ca

mp

an

ifo

rmes

0.1

547

8 ±

0.0

009

7

1.8

411

± 0

.00

33

62

.44 ±

0.4

9

9.5

28 ±

0.0

64

9

.41

2 ±

0.0

84

O-8

5

El

Cas

till

o

Over

lies

red

rect

ang

ula

r m

oti

f,

Ga

lerí

a d

el B

iosn

te

0.2

579

± 0

.00

64

2

.28

27

± 0

.00

65

3

.44

6 ±

0.0

92

1

2.9

5 ±

0.3

4

10

.1 ±

1.3

O-2

3

Tit

o B

ust

illo

Over

lies

red

vulv

a,

Cá

ma

ra d

e la

s

vulv

as

0.2

153

± 0

.00

16

1.6

909

± 0

.00

35

3

.02

1 ±

0.0

21

1

4.7

4 ±

0.1

2

11

.1 ±

1.7

O-9

7

La

Pas

iega

Over

lies

red

dee

r,

Pas

iega

C

0.2

940

± 0

.00

22

2

.60

04

± 0

.00

53

9

.17

7 ±

0.0

67

1

2.9

5 ±

0.1

0

11

.89 ±

0.4

5

O-1

7

Tit

o B

ust

illo

O

ver

lies

vio

let

ho

rse,

En

sem

ble

IX

0

.11

03

6 ±

0.0

006

1

0.8

73

1 ±

0.0

01

4

4.8

28 ±

0.0

22

1

4.8

03

± 0

.09

2

12

.5 ±

1.2

O-9

9

La

Pas

iega

Over

lies

red

do

t,

Pas

iega

C

0.3

894

± 0

.00

24

3

.46

01

± 0

.00

60

3

3.8

6 ±

0.1

6

12

.863

± 0

.08

5

12

.58 ±

0.1

4

O-4

0

Las

Ag

uas

O

ver

lies

red

and

0

.13

33

8 ±

0.0

007

0

1.1

321

± 0

.00

19

1

7.6

52

± 0

.07

5

13

.656

± 0

.08

0

13

.07 ±

0.3

0

engra

ved

bis

on,

Pri

nci

pal

Pan

el

O-1

4

Tit

o B

ust

illo

O

ver

lies

red

ho

rse,

En

sem

ble

X

0.2

424

± 0

.00

15

1

.65

78

± 0

.00

32

5

.12

8 ±

0.0

26

1

7.0

9 ±

0.1

2

14

.6 ±

1.1

O-8

6

El

Cas

till

o

Over

lies

bla

ck

bis

on,

El

Pa

so

0.5

580

± 0

.00

78

3

.66

35

± 0

.00

87

4

.93

1 ±

0.0

84

1

7.7

0 ±

0.2

7

15

.06 ±

0.9

9

O-1

2

Tit

o B

ust

illo

R

ed h

ors

e hea

d,

En

sem

ble

X

0.2

346

± 0

.00

17

1

.64

74

± 0

.00

35

9

.59

5 ±

0.0

63

1

6.6

1 ±

0.1

4

15

.33 ±

0.6

0

O-9

T

ito

Bust

illo

R

ed h

ors

e,

En

sem

ble

X

0.1

112

± 0

.00

10

0

.73

66

± 0

.00

18

9

.02

7 ±

0.0

88

1

8.0

5 ±

0.1

9

16

.55 ±

0.8

1

O-6

7

El

Cas

till

o

New

gro

wth

of

bro

ken

sca

rf

stal

agti

te w

ith r

ed

dis

k,

Ga

lerí

a d

el

Bis

on

te

0.2

174

± 0

.00

15

1

.42

05

± 0

.00

33

1

4.9

1 ±

0.1

3

18

.00 ±

0.1

4

17

.11 ±

0.4

4

O-8

1

El

Cas

till

o

Over

lies

red

dis

k,

Co

rred

or

Tec

ho

de

las

Ma

no

s

0.6

046

± 0

.00

44

3.7

396

± 0

.00

71

2

7.2

2 ±

0.2

5

18

.86 ±

0.1

5

18

.36 ±

0.2

3

O-7

2

La

Pas

iega

Over

lies

red

tria

ng

le,

Pas

iega

C

0.7

673

± 0

.00

33

4

.82

03

± 0

.00

90

2

60

.40

± 0

.64

1

8.5

19

± 0

.09

2

18

.468

± 0

.09

4

O-4

3

Las

Ag

uas

Over

lies

red

quad

rang

ula

r

sym

bo

l, C

ham

ber

of

Engra

vin

gs

0.2

257

± 0

.00

10

1

.34

94

± 0

.00

26

1

81

.0 ±

1.1

1

9.8

3 ±

0.1

0

19

.75 ±

0.1

1

O-5

3

Alt

amir

a

Over

lies

red

‘sp

ott

ed o

utl

ine’

ho

rse,

Tec

ho

de

los

Po

lícr

om

os

0.2

884

± 0

.00

13

1

.54

71

± 0

.00

26

1

07

.07

± 0

.20

2

2.2

6 ±

0.1

1

22

.11 ±

0.1

3

O-7

0

Las

Ag

uas

O

ver

lies

bro

wn ‘

T’

sig

n,

Pri

nci

pal

Pan

el

0.2

266

± 0

.00

13

1

.17

72

± 0

.00

21

1

8.0

3 ±

0.0

85

2

3.2

2 ±

0.1

6

22

.29 ±

0.4

7

O-8

0

El

Cas

till

o

Over

lies

bla

ck

ind

eter

min

ate

anim

al,

Co

rred

or

Tec

ho

de

las

Ma

no

s

0.7

879

± 0

.00

47

3

.98

28

± 0

.00

73

3

0.0

1 ±

0.1

5

23

.43 ±

0.1

6

22

.88 ±

0.2

7

O-5

8

El

Cas

till

o

Over

lies

red

neg

ativ

e han

d

sten

cil,

Tec

ho

de

las

Ma

no

s

0.5

272

± 0

.00

20

2

.57

74

± 0

.00

49

2

22

.70

± 0

.49

2

4.4

2 ±

0.1

1

24

.34 ±

0.1

2

O-2

1

Tit

o B

ust

illo

Red

pig

ment

asso

ciat

ed w

ith

anth

rop

om

orp

hic

figure

, G

ale

ría

de

los

An

tro

po

mo

rfo

s

0.6

252

± 0

.00

31

1

.80

38

± 0

.00

37

2

.17

± 0

.01

4

4.9

4 ±

0.2

9

29

.65 ±

0.5

5a

O-6

9

El

Cas

till

o

Red

dis

k,G

ale

ría

de

los

Dis

cos

0.7

512

± 0

.00

29

2

.70

72

± 0

.00

51

7

88

.24

± 5

.5

34

.28 ±

0.1

7

34

.25 ±

0.1

7

O-5

0

Alt

amir

a

Over

lies

red

clav

ifo

rm-l

ike

sym

bo

l, T

ech

o d

e

los

Po

lícr

om

os

0.4

933

± 0

.00

24

1.6

594

± 0

.00

30

1

7.4

73

± 0

.06

8

37

.60 ±

0.2

3

36

.16 ±

0.6

1

O-8

2

El

Cas

till

o

Over

lies

red

neg

ativ

e han

d

sten

cil

and

und

erli

es

yel

low

outl

ined

bis

on,

Pa

nel

de

las

Ma

no

s

0.5

111

5 ±

0.0

029

1

.69

70

± 0

.00

35

4

8.8

1 ±

0.4

9

38

.15 ±

0.2

7

37

.63 ±

0.3

4

O-8

3

El

Cas

till

o

Over

lies

lar

ge

red

dis

k,

Pa

nel

de

las

Ma

no

s

0.3

573

2 ±

0.0

022

1

.10

48

± 0

.00

20

2

8.6

4 ±

0.2

9

42

.38 ±

0.3

3

41

.40 ±

0.5

7

Tab

le S

1 R

esu

lts

of

the

U-s

erie

s d

iseq

uil

ibri

um

dat

ing o

f sa

mp

les

of

calc

ite

fro

m a

bove

the

art

(and,

thu

s, m

inim

um

ages

on

ly)

plo

tted

in

Fig

. 2

. Is

oto

pic

rati

os

are

giv

en a

s ac

tivit

y r

atio

s, e

rrors

are

at

2σ

. A

ges

are

corr

ecte

d f

or

det

ritu

s usi

ng a

n a

ssum

ed 2

32T

h/2

38T

h a

ctiv

ity o

f 1

.25

0 ±

0.6

25

an

d 2

30T

h/2

38U

an

d

23

4U

/23

8U

at

equ

ilib

riu

m, ex

cep

t (a

) w

hic

h i

s co

rrec

ted u

sing m

easu

red v

alues

on i

nso

lub

le r

esid

ue

23

0T

h/2

32T

h=

0.8

56

1 ±

0.0

03

9.

References and Notes

1. A. Sinclair, Archaeology: Art of the ancients. Nature 426, 774 (2003).

doi:10.1038/426774a Medline

2. J. Zilhão, The emergence of ornaments and art: An archaeological perspective on the

origins of behavioural “modernity”. J. Archaeol. Res. 15, 1 (2007).

doi:10.1007/s10814-006-9008-1

3. N. J. Conard, Cultural modernity: Consensus or conundrum? Proc. Natl. Acad. Sci.

U.S.A. 107, 7621 (2010). doi:10.1073/pnas.1001458107 Medline

4. J. Fortea, La plus ancienne production artistique du Paléolithique ibérique. In Pitture

Paleolitiche nelle Prealpi venete. Grotta di Fumane e Riparo Dalmeri, A.

Broglio, G. Dalmeri, Eds. (Museo Civico di Storia Naturale di Verona,Verona,

2005), pp. 89–99.

5. A. Leroi-Gourhan, B. Delluc, G. Delluc, Préhistoire de l’art occidental (Mazenod,

Paris, 1995)

6. P. B. Pettitt, A. W. G. Pike, Dating European cave art: Progress, prospects, problems.

J. Archaeol. Method Theory 14, 27 (2007). doi:10.1007/s10816-007-9026-4

7. F. J. Fortea Pérez, Trente-neuf dates C14-SMA pour l’art pariétal paléolithique des

Asturias. Bull. Soc. Préhistorique Ariège Pyrénées LVII, 7 (2002).

8. A. W. G. Pike et al., Verification of the age of the Palaeolithic rock art at Creswell

Crags, UK. J. Archaeol. Sci. 32, 1649 (2005). doi:10.1016/j.jas.2005.05.002

9. D. L. Hoffmann,230

Th isotope measurements of femtogram quantities for U-series

dating using multi ion counting (MIC) MC-ICPMS. Int. J. Mass Spectrom. 275,

75 (2008). doi:10.1016/j.ijms.2008.05.033

10. D. L. Hoffmann et al., Procedures for accurate U and Th isotope measurements by

high precision MC-ICPMS. Int. J. Mass Spectrom. 264, 97 (2007).

doi:10.1016/j.ijms.2007.03.020

11. Materials and methods are available as supplementary materials on Science Online.

12. P. J. Reimer et al., IntCal09 and Marine09 radiocarbon age calibration curves, 0–

50,000 years cal BP. Radiocarbon 51, 1111 (2009).

13. J. Zilhão, Chronostratigraphy of the Middle-to-Upper Palaeolithic transition in the

Iberian peninsula. Pyrenae 37, 7 (2006).

14. J. Zilhão et al., Pego do Diabo (Loures, Portugal): dating the emergence of

anatomical modernity in westernmost Eurasia. PLoS ONE 5, e8880 (2010).

doi:10.1371/journal.pone.0008880 Medline

15. J. Maroto et al., Current issues in late Middle Palaeolithic chronology: New

assessments from northern Iberia. Quat. Int. 247, 15 (2012).

doi:10.1016/j.quaint.2011.07.007

16. L. G. Straus, The Upper Palaeolithic of Cantabrian Spain. Evol. Anthropol. 14, 145

(2005). doi:10.1002/evan.20067

17. J. Zilhão, T. Aubry, F. Almeida, Un modèle technologique pour le passage du

Gravettien au Solutréen dans le Sud-Ouest de l'Europe. In Les faciès

leptolithiques du nord-ouest méditerranéen: milieux naturels et culturels. XXIVe

Congrès Préhistorique de France. Carcassonne, 26-30 Septembre 1994, D.

Sacchi, Ed. (Société Préhistorique Française, Paris, 1999), pp. 165–183.

18. C. Renard, Continuity or discontinuity in the Late Glacial Maximum of south-western

Europe: the formation of the Solutrean in France. World Archaeol. 43, 726

(2011). doi:10.1080/00438243.2011.624789

19. I. J. Fairchild, S. Frisia, A. Borsato, A. F. Tooth, Speleothems. In Geochemical

Sediments and Landscapes, D. J. Nash, S. J. McLaren, Eds. (Blackwell, Oxford,

2007), pp. 200–245.

20. A. Beltran, The Cave of Altamita (Harry Abrams, New York, 1999)

21. C. González-Sainz, R. Cacho Toca, R. Fukazawa, Arte Paleolítico en la Región

Cantábrica (Universidad de Cantabria, Cantabria, Spain, 2003).

22. F. Bernaldo de Quirós, Reflections on the art of the cave of Altamira. Proc.

Prehistoric Soc. 57, 81 (1991).

23. C. Heras Martín, R. Montes Barquín, J. A. Lasheras, P. Rasines, P. Fatás Monforte,

Nuevas dataciones de la cueva de Altamira y su implicación en la cronología de

su arte rupestre paleolítico. Cuadernos de Arte Rupestre de Moratalla 4, 117

(2008).

24. H. Breuil, E. Cartailhac, H. Obermaier, M. E. Boyle, The Cave of Altamira at

Santillana del Mar, Spain (Tip. de Archivos, Madrid, 1935).

25. H. Breuil, Quatre cents siècles d’art pariétal. Les cavernes ornées de l’age du renne

(Centre d’Etudes de Documentation Préhistoriques, Montignac, Paris, 1952).

26. M. Lorblanchet, Les Grottes Ornées de la Préhistoire. Nouveaux Regards (Editions

Errance, Paris, 1995).

27. F. Djindjian, L’art Paléolithique dans son système culturel: essais de corrélations. 1.

Chronologie, ‘styles’ et ‘cultures’. In Art Mobilier Paléolithique Supérieur en

Europe Occidental, A-C. Welté, E. Ladier, Eds. [Actes du Colloque 8.3, Congrès

de l’UISPP (Union Internationale des Sciences Préhistoriques et

Protohistoriques), Université de Liège, Liège, Belgium, 2004], pp. 249–256.

28. J. Clottes, J. Courtin, The Cave Beneath the Sea: Palaeolithic Images at Cosquer (H.

N. Adams, New York, 2006).

29. R. de Balbín, J. J. Alcolea, M. A. González, El macizo de Ardines, un lugar mayor

del arte paleolítico europeo. In El arte prehistórico desde los inicios del s. XXI.

Primer symposium internacional de arte prehistórico de Ribadesella, R. de

Balbín, P. Bueno, Eds. (Asociación Cultural de Amigos de Ribadesella,

Ribadesella, Spain, 2003).

30. J. Clottes, Return to Chauvet Cave: Excavating the Birthplace of Art (Thames and

Hudson, London, 2003).

31. P. B. Pettitt, P. Bahn, C. Züchner, The Chauvet conundrum: Are claims for the

‘birthplace of art’ premature? In An Enquiring Mind: Studies in Honor of

Alexander Marshack, P. Bahn, Ed. (American School of Prehistoric Research

Monograph Series, Oxbow Oxford, City, 2009), pp. 239–262.

32. A. Broglio, M. de Stefani, F. Gurioli, M. Peresani, The Aurignacian paintings of the

Fumane Cave (Lessini mountains, Venetian prealps). Int. Newsl. Rock Art 44, 1

(2006).

33. B. Delluc, G. Delluc, L’Art Pariétal Archaïque en Aquitaine (Gallia Préhistoire

XXVIII supplément, CNRS, Paris, 1991).

34. M. García-Diez, J. A. Mujika Alustiza, M. Sasieta, J. Arruabarrena, J. Alberdi,

Astigarraga Cave (Deba, Guipúzcoa, Spain) Archaeological work and human

occupation. Int. Newsl. Rock Art 60, 13 (2011).

35. H. Alcalde del Río, H. Breuil, L. Sierra, Les cavernes de la Région Cantabrique (A

Chene, Mónaco, 1911).

36. M. Ivanovich, R. S. Harmon Eds, Uranium-Series Disequilibrium: Applications to

Earth, Marine, and Environmental Sciences (Oxford Univ. Press, Oxford, ed. 2,

1992).

37. K. H. Wedepohl, The composition of the continental-crust. Geochim. Cosmochim.

Acta 59, 1217 (1995). doi:10.1016/0016-7037(95)00038-2

38. A. Kaufman, An evaluation of several methods for determining 230

Th/U ages in

impure carbonates. Geochim. Cosmochim. Acta 57, 2303 (1993).

doi:10.1016/0016-7037(93)90571-D

39. D. Scholz, D. L. Hoffmann,230

Th/U-dating of fossil corals and speleothems. Quat.

Sci. J. 57, 52 (2008).