Content uploaded by Anthony Dosseto
Author content
All content in this area was uploaded by Anthony Dosseto on Jun 19, 2015
Content may be subject to copyright.
1
Pleistocene cave art from Sulawesi, Indonesia
Authors: M. Aubert1,2†*, A. Brumm3,4*, M. Ramli5, T. Sutikna6,7, E. W. Saptomo7, B. Hakim8, M. J.
Morwood#, G. D. van den Bergh6, L. Kinsley9, A. Dosseto2,10
Affiliations:
1Place, Evolution and Rock Art Heritage Unit (PERAHU), Griffith University, Gold Coast, QLD 4222, Australia.
†E-mail: m.aubert@griffith.edu.au
2Wollongong Isotope Geochronology Laboratory, University of Wollongong, Wollongong, NSW 2522, Australia.
3Environmental Futures Research Institute, Griffith University, Brisbane, QLD 4111, Australia.
4School of Earth & Environmental Sciences, University of Wollongong, Wollongong, NSW 2522, Australia.
5Balai Pelestarian Peninggalan Purbakala, Makassar 90111, Indonesia.
6Centre for Archaeological Science, University of Wollongong, Wollongong, NSW 2522, Australia.
7National Centre for Archaeology (ARKENAS), Jakarta 12001, Indonesia.
8Balai Arkeologi Makassar, Makassar 90242, Indonesia.
9Research School of Earth Sciences, The Australian National University, Canberra, ACT 0200, Australia.
10GeoQuEST Research Centre, University of Wollongong, Wollongong, NSW 2522, Australia.
*These authors contributed equally
#Deceased
The authors declare no competing financial interests.
Archaeologists have long been puzzled by the appearance in Europe ~40 to 35 thousand years
(kyr) ago of a rich corpus of sophisticated artworks, including parietal art (i.e., paintings,
drawings and engravings on immobile rock surfaces)1,2 and portable art (e.g., carved
figurines)3,4, and the absence or scarcity of equivalent, well-dated evidence elsewhere,
especially along early human migration routes in South Asia and the Far East, including
Wallacea and Australia5-8, where modern humans (Homo sapiens) were established by 50 kyr
ago9,10. Here, using uranium-series dating of coralloid speleothems directly associated with 12
hand stencils and two figurative animal depictions from seven cave sites in the Maros karsts
of Sulawesi, we show that rock art traditions on this Indonesian island are at least compatible
in age with the oldest European art11. The earliest dated image from Maros, with a minimum
age of 39.9 kyr, now represents the oldest known hand stencil in the world. In addition, a
painting of a babirusa (‘pig-deer’) made at least 35.4 kyr ago is amongst the earliest dated
figurative depictions worldwide, if not the earliest one. Among the implications, it can now be
2
demonstrated that humans were producing rock art by ~40 kyr ago at opposite ends of the
Pleistocene Eurasian world.
Sulawesi is the world’s eleventh largest island and the biggest and probably oldest in Wallacea, the
zone of oceanic islands between continental Asia and Australia. The Eocene to middle Miocene
limestones of the Maros and Pangkep regions lie between 4°7´S and 5°1´S and cover an area of
~450 km2 parallel to the west coast of the island’s southwestern peninsula12 (Fig. 1). Rivers draining
the volcanic highlands to the east cut down into the basal limestone, forming clusters of plateau-like
karst towers that rise abruptly from the surrounding alluvial plains12. Extensive networks of
footcaves formed around the tower bases and now harbor abundant evidence of prehistoric human
occupation13. Cemented breccia banks containing archaeological material occur on the rear walls of
many caves and rockshelters14,15, and at least 90 rock art sites are recorded. While multiple cave and
shelter sites have been excavated since the 1930s16, only two with Pleistocene sequences, Leang
Burung 2 (ref. 13) and Leang Sakapao 1 (ref. 17), are thus far reported (Fig. 1). The oldest, Leang
Burung 2, a cliff-foot shelter with a minimum age for the excavated deposits of 31,260 ± 320
radiocarbon years BP (35,248 ± 420 calendar years BP)13, previously provided the earliest dated
evidence for humans on Sulawesi. The Pleistocene deposits from both sites yielded evidence of
pigment use in the form of facetted haematite nodules13 and ochre-smeared stone tools17.
The Maros-Pangkep rock art was first recorded in the 1950s15 and has been extensively
studied by Indonesian researchers, although few detailed reports are published. On the basis of
superimposition, two broad periods of prehistoric art production are defined18. The earliest of these
is characterized by human hand stencils (made by spraying wet pigment around hands pressed
against rock surfaces) and, less commonly, large naturalistic paintings of endemic Sulawesi land
mammals, including the dwarfed bovid anoa (Anoa sp.), Celebes warty pig (Sus celebensis) and the
‘pig-deer’ babirusa (Babyrousa sp.). These wild animal species are most commonly depicted in
profile as irregularly infilled outlines18.
The later rock art phase in the Maros-Pangkep karsts lacks images of this nature. It is instead
typified by small depictions of zoomorphs (including dogs and other domesticated species),
anthropomorphs, and a wide range of geometric signs, most commonly drawn onto rock surfaces
using black pigment (possibly charcoal)18. This art can plausibly be attributed to early Austronesian
immigrants on the basis of stylistic elements19, and is thus a few thousand years old at most20.
The red- and mulberry-coloured motifs of the earlier phase typically occur on high roofs,
elevated parts of rock walls or other difficult-to-access areas in caves and shelters18. They are
located both close to site entrances and within deep, dark chambers and passages. In most cases the
art is poorly preserved, surviving only as weathered patches of pigment on exfoliated rock surfaces.
3
At some sites, better-preserved art is partly or almost completely obscured by dense clusters of
small coralloid speleothems (‘cave popcorn’) up to ~10 mm thick, which form when thin films of
water precipitate on rock surfaces21. At one Maros cave site, Leang Bulu Bettue (Fig. 1), we
observed Austronesian style drawings on a ‘fresh’ limestone ceiling formed by shedding of an
earlier surface containing faded hand stencils (Extended Data Fig. 1), suggesting that even in recent
prehistoric times this art was in an advanced state of deterioration. However, local custodians report
that the loss of the art has accelerated in recent decades.
To determine the age of the earliest rock art in the Maros karsts we undertook an extensive
program of uranium-series dating of coralloid speleothems directly associated with the motifs. The
sampled materials all comprise static coralloids that formed directly on top of clearly discernable
motifs, offering the possibility to obtain minimum ages for the underlying rock art. In some cases,
hand stencils and paintings were made over coralloids that then continued to grow, providing an
opportunity to obtain both minimum and maximum ages for the art.
We collected a total of 19 coralloid samples associated with 14 individual motifs (12 hand
stencils and two figurative animal depictions) (Figs. 2-3 and Extended Data Figs. 2-9) at seven cave
sites in the Maros karsts (Fig. 1). Six of these sites are located within a ~1 km radius in the
Bantimurung region, close to Leang Burung 2. Four of the Bantimurung sites (Gua Jing, Leang
Barugayya 1 and 2, and Leang Timpuseng) are situated in a large limestone outlier roughly 2 km in
diameter and 180 m high12. Leang Sampeang is located in an elevated niche on tall limestone cliffs
~500 m east of the outlier, while Leang Lompoa occurs at the base of an adjacent karst inselberg.
The seventh cave site, Leang Jarie, is in the Simbang district southeast of Bantimurung (Fig. 1).
To provide an internal check of the microstratigraphic order of ages we took a minimum of
three (and up to six) aliquots from every sample (except for Samples Leang Jarie 1 and 2 [2012]),
one from under the pigment layer, and two or more from above it, giving a total of 55 uranium-
series age determinations (Supplementary Information). In addition, at Leang Jarie (Fig. 3), Leang
Barugayya 2 (Extended Data Fig. 6) and Leang Sampeang (Extended Data Fig. 9) we dated two
coralloids that had formed over the same motif. At Leang Lompoa (Extended Data Fig. 3) and
Leang Jarie (Extended Data Fig. 2) we also dated two samples taken from different parts of the
same coralloid. Dating results for these five sets of paired samples are internally consistent
(Supplementary Information), demonstrating the robustness of the ages for the associated motifs.
Minimum ages for the Maros rock art motifs (n = 14) span the time range between 39.9 and
17.4 kyr ago, with the majority dating to more than 25 kyr ago (Table 1 and Supplementary
Information). The oldest dated motif is a hand stencil from Leang Timpuseng, which has a
minimum age of 39.9 kyr (Fig. 2), and now represents the earliest evidence for humans on Sulawesi,
as well as the oldest known example of this widespread art form. This motif is located on a 4-m-
4
high ceiling next to a large irregularly infilled painting of a female babirusa, which has a minimum
age of 35.4 kyr (Fig. 2). At nearby Leang Barugayya 2, a large painting of an indeterminate animal
(probably a pig) has a minimum age of 35.7 kyr (Extended Data Fig. 6). The next oldest motif in
our assemblage is another hand stencil at Leang Jarie, which dates to at least 39.4 kyr ago (Fig. 3).
With the Leang Timpuseng hand stencil, and for many other motifs in our sample,
subsamples taken from below the pigment layer were more than 100 kyr in age (Supplementary
Information). These early dates represent calcium carbonate deposits (flowstone layers) present on
the rock face before the art was produced. At Gua Jing, we dated two distinct hand stencils, one of
which yielded minimum and maximum ages of 22.9 and 27.2 kyr, respectively (Extended Data Fig.
8). Thus, given that the Leang Timpuseng hand stencil has a minimum age of 39.9 kyr, we can infer
the existence in the Maros karsts of an artistic culture with a duration of at least ~13 kyr.
The discovery of rock art dating back at least 40 kyr ago on Sulawesi has important
implications for our understanding of the time-depth of early symbolic traditions in the region,
about which little is presently known. For instance, rock art complexes focused on hand stencils and
large animal paintings occur in the Bone karsts ~35 km east of Maros (Fig. 1), as well as west of
Sulawesi in Kalimantan (Borneo)22,23 and further afield in mainland Southeast Asia24. The northern
Australian rock art provinces of Arnhem Land25 and the Kimberley26 also display early art phases
(based on order of superimposition) characterized by hand stencils and large irregularly infilled
paintings of animals, including apparent images of extinct megafauna25,26, that are markedly similar
in style to the Maros art. Given that the deepest excavated deposits in northern Australia (dated to
~50-40 kyr ago) contain use-worn haematite crayons and other evidence of ochre processing and
use9,10,27, it is possible that an extensive archive of rock art may yet survive from the initial modern
human colonization of Australia and Southeast Asia.
There are also implications for the debate about the origins of Palaeolithic rock art per se,
which is invariably dominated by European data and for which there are two widely debated
models11,28. The first of these is that rock art originated in Europe and developed gradually over
thousands of years, beginning with abstract, nonfigurative imagery (e.g., geometric patterns) and
culminating in sophisticated naturalistic representations of animals, such as those in Altamira and
Lascaux dated to ~20 kyr ago11,28,29 and other late Upper Palaeolithic cave sites in western Europe.
This long-standing notion is given new impetus by recent uranium-series dating of rock art motifs
from 11 caves in northern Spain, which suggests that Europe’s earliest cave art was nonfigurative in
nature and that animal paintings do not appear until considerably later11,28. Currently, the oldest
dated rock art motif in Europe (and the world) is from El Castillo, where a single thin calcite
deposit overlying a red ‘disk’ yielded a minimum uranium-series age of 40.8 kyr11. The alternative
model is that cave art first appeared in Europe in fully developed form, as implied by the great
5
antiquity of the elaborate animal paintings from Chauvet Cave in southern France29. Although the
early chronology for this art is disputed30, the oldest animal image from Chauvet Cave is attributed
an age of 32,410 ± 720 radiocarbon years BP (~35,000 calendar years BP) on the basis of 14C-
dating of charcoal pigment29.
Our dating results from Sulawesi suggest that figurative art was already part of the cultural
repertoire of the first modern human populations to reach this region more than 40 kyr ago. It is
possible that rock art emerged independently at around the same time and at approximately both
ends of the spatial distribution of early modern humans. An alternative scenario, however, is that
cave painting was widely practiced by the first H. sapiens to leave Africa tens of thousands of years
earlier, and thus that naturalistic animal art from Leang Timpuseng and Leang Barugayya 2, as well
as Chauvet Cave in France, may well have much deeper origins outside both western Europe and
Sulawesi. If so, we may anticipate future discoveries of depictions of human hands, figurative art,
and other forms of image-making dating to the earliest period of the global dispersal of our species.
6
References
1. Lewis-Williams, D. J. The Mind in the Cave: Consciousness and the Origins Of Art (Thames &
Hudson, 2002).
2. White, R. et al. Context and dating of Aurignacian vulvar representations from Abri Castanet,
France. Proc. Natl. Acad. Sci. U.S.A. 109, 8450−8455 (2012).
3. Conard, N. J. Palaeolithic ivory sculptures from southwestern Germany and the origins of
figurative art. Nature 426, 830−832 (2003).
4. Dowson, T. A. & Porr, M. in The Archaeology of Shamanism (ed Price, N.) 165−177 (Routledge,
2001).
5. Aubert, M. A review of rock art dating in the Kimberley, Western Australia. J. Arch. Sci. 39,
573−577 (2012).
6. Brumm, A. & Moore, M. W. Symbolic revolutions and the Australian archaeological record.
Cambridge Archaeol. J. 15, 157−175 (2005).
7. Langley, M. C., Clarkson, C. & Ulm, S. From small holes to grand narratives: the impact of
taphonomy and sample size on the modernity debate in Australia and New Guinea. J. Hum. Evol.
61, 197−208 (2011).
8. Mellars, P. Going east: new genetic and archaeological perspectives on the modern human
colonization of Eurasia. Science 313, 796−800 (2006).
7
9. Roberts, R. G., Jones, R. & Smith, M. Thermoluminescence dating of a 50,000-year-old human
occupation site in northern Australia. Nature 345, 153−156 (1990).
10. Roberts, R. G. et al. The human colonisation of Australia: optical dates of 53,000 and 60,000
years bracket human arrival at Deaf Adder Gorge, Northern Territory. Quatern. Sci. Rev. 13,
575−583 (1994).
11. Pike, A. W. G. et al. U-series dating of Paleolithic art in 11 caves in Spain. Science 336,
1409−1413 (2012).
12. McDonald, R. C. Limestone morphology in South Sulawesi, Indonesia. Zeitschrift fur
Geomorphologie, suppl Bd. 26, 79−91 (1976).
13. Glover, I. C. Leang Burung 2: an Upper Palaeolithic rock shelter in south Sulawesi, Indonesia.
Mod. Quat. Re. S.E. Asia, 6, 1−38 (1981).
14. Glover, I. C. The effects of sink action on archaeological deposits in caves: an Indonesian
example. World Archaeol. 10, 302−317 (1979).
15. van Heekeren, H. R. Rock-paintings and other prehistoric discoveries near Maros (South West
Celebes). Laporan Tahunan Dinas Purbakala 1950, 22−35 (1952).
16. Bulbeck, D., Pasqua, M. & Di Lello, A. Culture history of the Toalean of South Sulawesi,
Indonesia. Asian Perspec. 39, 71−108 (2000).
8
17. Bulbeck, D., Sumantri, I. & Hiscock, P. Leang Sakapao 1, a second dated Pleistocene site from
South Sulawesi, Indonesia. Mod. Quat. Re. S.E. Asia, 18, 111−128 (2004).
18. Eriawati, Y. Lukisan di Gua-Gua Karst Maros-Pangkep, Sulawesi Selatan: Gambaran
Penghuni dan Matapencahariannya (Indonesian Ministry of Cultural Media Development, 2003).
19. O’Connor, S. Nine new painted rock art sites from East Timor in the context of the Western
Pacific region. Asian Perspect. 42, 96−128 (2003).
20. Simanjuntak, T. Austronesian in Sulawesi (Center for Prehistoric and Austronesian Studies,
2008).
21. Hill, C. A. & Forti, P. Cave Minerals of the World (National Speleological Society, 1997).
22. Fage, L. H. & Chazine, J. M. Bornéo la Mémoire des Grottes (Fage éditions, 2009).
23. Plagnes, V. et al. Cross dating (Th/U-14C) of calcite covering prehistoric paintings in Borneo.
Quat. Res. 60, 172−179 (2003).
24. Taçon, P. S. C. & Tan, N. H. in Rock Art News of the World 4 (eds Bahn, P., Franklin, N. &
Strecker, M.) 207−214 (Oxbow Books, 2012).
25. Chaloupka, G. Journey in Time, the World’s Longest Continuing Art Tradition: The 50,000-
year Story of the Australian Aboriginal Rock Art of Arnhem Land (Reed, 1993).
9
26. Morwood, M. J. Visions from the Past: The Archaeology of Australian Aboriginal Art (Allen &
Unwin, 2002).
27. David, B. et al. How old are Australia's pictographs? A review of rock art dating. J. Arch. Sci.
40, 3−10 (2013).
28. García-Diez, M. et al. Uranium series dating reveals a long sequence of rock art at Altamira
Cave (Santillana del Mar, Cantabria). J. Arch. Sci. 40, 4098−4106 (2013).
29. Valladas, H. et al. Bilan des datations carbone 14 effectuées sur des charbons de bois de la
grotte Chauvet. Bulletin de la Société Préhistorique Française 102, 109−113 (2005).
30. Combier, J. & Jouve, G. Nouvelles recherches sur l’identité culturelle et stylistique de la grotte
Chauvet et sur sa datation par la méthode du 14C. L’anthropologie (2014). Pre-print at
http://dx.doi.org/10.1016/j.anthro.2013.12.001
Acknowledgements This research was supported by grants from the Australian Research Council
to M.A. (DP110102898/DE140100254) and A.B. (DP0879624/DE130101560) and the Centre for
Archaeological Science (CAS), University of Wollongong (UOW). The fieldwork was authorized
by the director of the Makassar Heritage Department (B.P.P.P.), M. Said, and the director of the
National Centre for Archaeology in Jakarta (ARKENAS), B. Sulistyanto. We further acknowledge
Balai Arkeologi Makassar, the Indonesian State Ministry of Research and Technology, and the
Geological Survey Institute in Bandung, for facilitating the research. We thank UOW’s Deputy
Vice Chancellor (Research), J. Raper, for additional project support. Field assistants included M.
Andi Pampang and A. A. Oktaviana. Technical laboratory assistance involved G. Mortimer, H.
10
Price, L. Sweetman and L. Yu., while C. Owers provided map data. P. Taçon and M.W. Moore are
thanked for critical feedback on the manuscript.
Author Contributions A.B. and T.S. conceived the study with M.A., as part of a wider project led
by M.R., E.W.S. and B.H., in collaboration with A.B., M.J.M. and G. v.d. B.. M.A. and A.B.
identified the samples. M.A collected the samples and conducted the uranium-series dating with
A.D.. M.A. and A.B. wrote the manuscript.
Methods Summary A small segment (~100-200 mm2) of each coralloid was removed from the
rock art panels using a battery operated rotary tool equipped with a diamond saw blade. Each
coralloid sample was sawn in situ so as to produce a continuous microstratigraphic profile
extending from the outer surface of the coralloid through the pigment layer and into the underlying
rock face. The only exception was Leang Jarie 1 and 2 (2012), which were sawn in situ but not
through the pigment layer. All of the sampled coralloids comprised multiple layers of dense and
nonporous calcite. The identification of a pigment layer overlaid by an extensive accumulation of
calcite laminations within each coralloid (except for Leang Jarie 1 and 2 [2012]) unambiguously
demonstrates that the sampled speleothems formed over the motifs (see Figs. 2-3 and Extended
Data Figs. 2-9). In the laboratory, the samples were microexcavated in arbitrary ‘spits’ over the
entire surface of the coralloids, creating a series of aliquots less than 1-mm-thick. The pigment layer
was visible across the entire length of the sample (except for Leang Jarie 1 and 2 [2012]). In total,
we obtained 55 uranium-series age determinations (an additional two failed to produce enough
signal for age determination) (Table 1 and Supplementary Information). The uranium-series
isotopes were measured on a ThermoFinnigan Neptune Plus Multi-Collector MC-ICP-MS and
calculation of ages and initial 234U/238U ratios was done with Isoplot 3.75. Corrections for detrital
components were calculated assuming the bulk Earth 232Th/238U concentration ratio of the upper
crust of 3.8 ± 50% and secular equilibrium for 230Th, 234U and 238U. In the main text, minimum ages
11
are quoted as measured age minus 2σ and maximum ages as measured age plus 2σ rounded to 1
decimal place.
31. Bourdon, B., Henderson, G. M., Lundstrom, C. C. & Turner, S. P. Uranium-series
Geochemistry (Mineralogical Society of America, 2003).
32. McCulloch M. T. & Esat, T. The coral record of last interglacial sea levels and sea surface
temperatures. Chem. Geol. 169, 107−129 (2000).
33. McCulloch, M. T. & Mortimer, G. E. Applications of the 238U–230Th decay series to dating of
fossil and modern corals using MC-ICPMS. Aust. J. Earth Sci. 55, 955−965 (2008).
34. Cheng, H., Adkins, J., Edwards, R. & Boyle, E. U-Th dating of deep-sea corals. Geochim.
Cosmochim. Ac. 64, 2401–2416 (2000).
35. Leclerc, P. in Expédition Thaï-Maros 86, 147-153 (Association Pyreneenne de Speleologie,
1987).
12
Figure Legends
Figure 1 | Location of the study area. a, Sulawesi is situated east of Borneo in the Wallacean
archipelago; b, The Maros-Pangkep karsts (shown as area of high relief) are near the town of Maros
on Sulawesi’s southwestern peninsula. The separate karst region of Bone is further east; c,
Archaeological sites included in this study: 1) Leang Barugayya 2; 2) Leang Barugayya 1; 3) Gua
Jing; 4) Leang Bulu Bettue; 5) Leang Sampeang; 6) Leang Timpuseng; 7) Leang Burung 2; 8)
Leang Lompoa; and 9) Leang Jarie. Gua Jing and Leang Barugayya 1 and 2 are separate cave sites
interconnected by a system of phreatic passages. Map data: a, b, Copyright ©ESRI (2008); c,
Copyright ©2014 Google Maps.
Figure 2 | Dated rock art from Leang Timpuseng. a, b, Locations of the dated coralloid
speleothems and associated paintings, a hand stencil and a large naturalistic depiction of an animal
shown in profile. Although the animal figure is badly deteriorated and obscured by coralloids, we
interpret it as a female babirusa. A painted red line below the babirusa (not clearly visible in a, but
illustrated in b) appears to represent the ground surface on which the animal is standing or walking.
The rock art panel is located on the ceiling about 8 m from the cave entrance and 4 m above the
current cave floor; c, d, Profiles of the coralloid speleothems showing the microexcavated
subsamples bracketing the age of the paintings. We interpret the similar ages for the overlying
aliquots as a result of fast growing speleothems. Tracing credit: Leslie Refine “Graph & Co”
(France).
Figure 3 | Dated rock art from Leang Jarie. a, b, Locations of the dated coralloid speleothems
and associated hand stencils. The hand stencils are part of a 4-m-long art panel located in a dark
recess along the eastern wall of the cave, about 5 m from the entrance and 1.5 m above the floor. c,
Profile of the coralloid speleothem (Leang Jarie 1, 2013) showing the microexcavated subsamples
13
bracketing the age of the paintings. The Leang Jarie 1 (2012) sample is from above the pigment
layer and so only provides a minimum age for the underlying hand stencils. Tracing credit: Leslie
Refine “Graph & Co” (France).
Table title
Table 1 | Results of uranium-series disequilibrium dating showing the minimum age of each
dated rock art motif (all isotopic ratios are activity ratios; errors are at 2s).
Sample Site Description 230Th/238U 234U/238U 230Th/232Th Uncorrected
Age (kyr) +2σ (kyr) -2σ
(kyr)
Corrected
Age (kyr) +2σ (kyr) -2σ (kyr) Initial 234U/238U
LL3.2 Leang
Lompoa Overlies hand stencil 0.1525 ± 0.0022 1.0067 ± 0.0014 137 17.87 0.27 0.28 17.77 0.42 0.42
1.0070 ± 0.0014
LB2.3 Leang
Barugayya 1 Overlies hand stencil 0.1624 ± 0.0077 0.9812 ± 0.0027 858 20.00 1.00 1.00 19.70 1.00 1.00
0.9801 ± 0.0028
LB3.3 Leang
Barugayya 1 Overlies hand stencil 0.2004 ± 0.0214 0.9799 ± 0.0025 428 24.90 2.90 2.90 24.90 3.10 3.00
0.9784 ± 0.0026
GJ2.2 Gua Jing Overlies hand stencil 0.1996 ± 0.0044 0.9943 ± 0.0009 50 24.40 0.60 0.59 24.00 1.10 1.10
0.9939 ± 0.0009
LB1.2 Leang
Barugayya 1 Overlies hand stencil 0.2308 ± 0.0211 0.9831 ± 0.0025 360 29.10 3.00 2.90 29.10 3.20 3.10
0.9817 ± 0.0028
LL1.3 Leang
Lompoa Overlies hand stencil 0.2322 ± 0.0030 1.0128 ± 0.0024 121 28.31 0.44 0.43 28.10 0.66 0.67
1.0138 ± 0.0025
LL2.2 Leang
Lompoa Overlies hand stencil 0.2391 ± 0.0064 1.0065 ± 0.0007 133 29.50 0.92 0.89 29.30 1.20 1.10
1.0070 ± 0.0008
GJ1.3 Gua Jin
g
Sequence of aliquots 0.2525 ± 0.0048 0.9998 ± 0.0010 31 31.70 0.69 0.69 30.90 1.70 1.80
0.9998 ± 0.0011
LS1.2 Leang
Sampeang Overlies hand stencil 0. 2549 ± 0.0044 0.9823 ± 0.0007 324 32.70 0.66 0.65 32.60 0.76 0.76
0.9806 ± 0.0007
LJ2 Lean
g
Jarie Overlies hand st encil 0.2738 ± 0.0022 0.9942 ± 0.0010 422 35.04 0.32 0.32 34.98 0.41 0.41
0.9935 ± 0.0011
LT1.2 Leang
Timpuseng
Overlies barbirusa
painting 0.2927 ± 0.0100 1.0163 ± 0.0023 682 37.00 1.50 1.50 36.90 1.60 1.50
1.0181 ± 0.0025
LB4.2 Leang
Barugayya 2
Overlies
undetermined animal
fi
g
ure
0.3481 ± 0.0385 1.0080 ± 0.0042 18 46.00 6.40 6.20 44.00 9.10 8.30
1.0091 ± 0.0046
LJ1 Leang Jarie Overlies hand stencil 0.3006 ± 0.0018 0.9839 ± 0.0014 1474 39.69 0.29 0.30 39.67 0.32 0.32
0.9820 ± 0.0015
LT2.3 Leang
Timpuseng Overlies hand stencil 0.3177 ± 0.0055 1. 0156 ± 0.0011 2845 40.80 0.83 0.83 40.70 0.87 0.84
1.0175 ± 0.0013
14
Supplementary Information
Supplementary Methods
Coralloid speleothems can be nodular, globular, botryoidal or coral-like in morphology. They form
from thin films of water precipitating on cave surfaces, resulting in concentric growth rings21. When
precipitated from saturated solutions, calcium carbonate usually contains small amounts of soluble
uranium (238U and 234U), which eventually decay to 230Th. The latter is essentially insoluble in cave
waters and will not precipitate with the calcium carbonate. This produces disequilibrium in the
decay chain where all isotopes in the series are no longer decaying at the same rate. Subsequently,
238U and 234U decay to 230Th until secular equilibrium is reached. Because the decay rates are
known, the precise measurement of these isotopes allows calculation of the age of the carbonate
formation31.
It is also common for secondary calcium carbonate to be contaminated by detrital materials, such as
wind-blown or waterborne sediments, and as such can lead to uranium-series ages that are
erroneously older than the true age of the sample. This is because the detrital fraction will
contribute to the overall amount of uranium-series nuclides so that the sample does not reflect a
radioactive disequilibrium related to the time of carbonate formation. The effects of detrital
contamination can be identified and often corrected for by measuring the activity of 232Th that is
solely present in the detrital fraction but which plays no part in the decay chain of uranium. An
indication of the degree of detrital contamination is expressed as 230Th/232Th activity, with high
values (>20) indicating little or no effect on the calculated age and low values (<20) indicating the
correction on the age will be significant31. Except for two samples (LL3.1 and B4.2), all our
samples have 230Th/232Th activity >20 indicating sample purity.
The small calcium carbonate samples were separately weighed in Savillex PFA vials. The samples
were covered with MilliQ water, and drops of Merck Ultrapur 60% HNO3 were added until
complete dissolution was achieved. A spike solution enriched in 236U-229Th was subsequently added
and the mixture was left to equilibrate overnight. The solutions were evaporated to dryness and then
redissolved in 1.5 M HNO3 ready for ion exchange chromatography, consisting of 0.25 mL of
Eichrom TRU resin over 0.1 mL of Eichrom pre-filter resin. The resins were cleaned by passing 3
M HCl, 0.2 M HCl and a 0.1 M HCl + 0.3 M HF mixture through the columns prior to use and then
preconditioned with 1.5 M HNO3. After the sample solutions were loaded onto the TRU resin bed
as solutions in 1.5 M HNO3, the columns were washed with 1.5 M HNO3 and 3 M HCl. Uranium
and thorium were imperfectly separated from the ion exchange medium using 0.2 M HCl (for
15
thorium), and 0.1 M HCl + 0.3 M HF (for uranium). Finally, the samples were evaporated to
dryness and redissolved in 4 ml 2% HNO3.
The U and Th solutions were introduced separately into a ThermoFinnigan Neptune Plus Multi-
Collector MC-ICP-MS equipped with a large interface pump, Jet Sample and Skimmer cones,
electrostatic analyzer, secondary electron multiplier (SEM), and retarding potential quadrupole
(RPQ) for high abundance sensitivity. Samples were aspirated using an ESI PFA-ST Aridus II
nebulizer at an uptake rate of ~0.1 mL/min. Sweep gas (Ar) flow rate was set to ~3-4 L/min and
Nitrogen was set to ~2-4 mL/min. Sensitivity was >1 V/ppb U.
Uranium isotopes were measured with the RPQ off, while thorium isotopes were measured with the
RPQ on. Isotopic ratios were corrected for background, tailing of 238U on 236U and 234U,
SEM/Faraday yield and instrumental mass bias (using 238U/235U = 137.88) after subtracting the
minor spike component. The SEM/Faraday yield was calculated externally using the NBS 960
standard by alternating 235U between the SEM and Faraday array while simultaneously measuring
238U on the Faraday array. This value was corrected for instrumental mass bias and compared with
the true value in SRM 960 = 0.007265. The SRM 960 standard was measured every two samples.
Relative gains derived from standard measurements were then interpolated to the unknowns. Other
standards used in this study are: AC-1, an Australian National University (ANU) coral powder with
a measured TIMS U-series age of 125 550 years32; and HU-1, a solution of secular equilibrium
Harwell Uraninite, also supplied by the ANU. AC-1 and HU-1 results are shown in Supplementary
Information, and in both cases are within error of the expected values. Total procedure blanks were
in the order of 0.9 pg for Th and 0.1 pg for U. Further details on our MC-ICP-MS procedure can be
found in ref. 33.
Calculation of ages and initial 234U/238U ratios was done using Isoplot 3.75 using the following
decay constants (dc) and half-lives (hl): 238Udc = 1.55125E-10; 238Uhl = 4.46831E+9, 234Udc =
2.82207E-6; 234Uhl = 2.45617E+5, 232Thdc = 4.94752E-11; 232Thhf = 1.401E+10, 230Thdc = 9.17052E-
6; 230Thhl = 7.55843E+4. Errors were calculated by Monte Carlo simulation (5,000 trials) ignoring
the uncertainties in the 235U and 238U decay constants. Corrections for detrital components were
calculated assuming the bulk Earth 232Th/238U concentration ratio of the upper crust of 3.8 ± 50%34
and secular equilibrium for 230Th, 234U and 238U.
16
Extended Data Figure 1 | Rock art panel on the ceiling at Leang Bulu Bettue. a, Black
drawings of early Austronesian style were made on a relatively freshly exposed limestone surface
and are superimposed over remnant patches of a much older surface, now extremely heavily
weathered and almost completely exfoliated, containing faded hand stencils (shown more clearly
and highlighted by arrows in b). The same rock art panel was documented and illustrated in a
publication by a team of French cavers in 1986, but the hand stencils were not identified35.
Extended Data Figure 2 | Dated rock art from Leang Jarie. a, Locations of the sampled
coralloid speleothems and associated hand stencils; b, Profile of the coralloid speleothem showing
the microexcavated subsamples bracketing the age of the paintings. The Leang Jarie 2 (2012)
sample is from above the pigment layer and so only provides a minimum age for the underlying
hand stencils.
Extended Data Figure 3 | Dated rock art from Leang Lompoa. a-c, Locations of the sampled
coralloid speleothems and associated hand stencil. Although heavily obscured by coralloid
speleothems, we interpret this image as a ‘mutilated hand’ stencil, which shows in outline a human
hand with two amputated digits or with the third and fourth fingers folded into the palm. The hand
stencil is located on the ceiling of a narrow, dimly lit passage leading off from the main entrance to
the cave. Samples Leang Lompoa 1 (2012) and Leang Lompoa 1 (2013) are part of the same cluster
of coralloid speleothems that formed over the hand stencil; d, e, Profile of the coralloid speleothems
showing the microexcavated subsamples bracketing the age of the motif. Note that sample LL1.2
(2012) does not represent the age of the hand stencil. The resultant age reflects a mixture of calcium
carbonate from below and above the pigment layer. Tracing credit: Leslie Refine “Graph & Co”
(France).
17
Extended Data Figure 4 | Dated rock art from Leang Lompoa. a, Locations of the sampled
coralloid speleothems and associated hand stencils. The hand stencils occur on a 2.5-m-high ceiling
in a small, dimly lit side chamber leading off from the cave mouth. The stencil at left (Leang
Lompoa 3) is stylistically distinct from the adjacent stencil (Leang Lompoa 2), with the fingers
modified by brushwork subsequent to stenciling to produce slender and pointy forms; b, c, Profiles
of the coralloid speleothems showing the microexcavated subsamples bracketing the age of the
hand stencils.
Extended Data Figure 5 | Dated rock art from Leang Barugayya 1. a, Locations of the sampled
coralloid speleothems and associated cluster of hand stencils. The hand stencils are situated on a
small rock art panel near the ceiling and close to the cave entrance. Samples LB1 and LB2 come
from two distinct hand stencils that are dark mulberry (almost black) in colour. Sample LB3 is from
over an adjacent red hand stencil; b-d, Profiles of the coralloid speleothems showing the
microexcavated subsamples bracketing the age of the hand stencils.
Extended Data Figure 6 | Dated animal painting from Leang Barugayya 2. a, b, Composite of
photographs showing the locations of the sampled coralloid speleothems and associated large
infilled red painting of an animal. Field photographs were altered in the software program DStretch
to enhance the image (b). The animal species depicted is unidentified due to the extent of
weathering and deterioration of the painting and the thick accumulation of coralloids over the art;
however, the painting appears to show in profile a large land mammal, probably a pig (a babirusa or
Sus celebensis), with the head facing right and the hindquarters at left; c, d, Profile of the coralloid
speleothems showing the microexcavated subsamples bracketing the age of the painting. Images a
and b courtesy of A. A. Oktaviana.
18
Extended Data Figure 7 | Dated rock art from Gua Jing. a, Location of the sampled coralloid
speleothem and associated hand stencil. The hand stencil is located on a stalactite curtain 15 m from
the cave entrance and 2 m above the current cave floor. The cave itself comprises a dark, winding
phreatic tube containing an extensive gallery of hand stencils and figurative animal motifs; b,
Profile of the coralloid speleothem showing the microexcavated subsamples bracketing the age of
the hand stencil.
Extended Data Figure 8 | Dated rock art from Gua Jing. a, Location of the sampled coralloid
speleothem and associated hand stencil; b, Profile of the coralloid speleothem showing the
microexcavated subsamples bracketing the age of the hand stencil.
Extended Data Figure 9 | Dated rock art from Leang Sampeang. a, Locations of the sampled
coralloid speleothems and associated hand stencil. Leang Sampeang is a 20-m-deep, narrow
chamber with paintings located on the ceiling at the back of the cave in complete darkness. In this
area, the cave is only 2.5 m wide and requires crawling to reach. Samples Leang Sampeang 1 and
Leang Sampeang 2 came from the same red hand stencil located 17 m from the cave entrance and
18 cm above the current cave floor; b, c, Profile of the coralloid speleothem showing the
microexcavated subsamples bracketing the age of the hand stencil.
