Narrative cave art in Indonesia by 51,200 years ago

Rationale for development and application of LA-U-series imaging

The use of the U-series disequilibrium method for rock art dating is relatively new and has drawn criticism from some authorities, especially in relation to the extremely old ages obtained for some European rock art and its attribution to Neanderthals. In essence, the U-series disequilibrium method relies on the dating of calcium carbonate material present on the surface of rock art, providing a minimum age for the parietal imagery on which it formed. Four main principles need to be met for the valid use of this approach. (1) The associated motif needs to have an anthropomorphic origin; that is, there must be no doubt that the imagery associated with the calcium carbonate is an artwork produced by humans rather than some kind of natural marking. (2) The relationship between the rock art and the associated calcium carbonate must be unambiguous. (3) The dated samples must be relatively ‘pure’ (that is, the calcium carbonate must not be too contaminated by detrital material, which renders it ‘dirty’). (4) It needs to be empirically demonstrated that the dated calcium carbonate samples have remained a closed-system for uranium and thorium, and, especially, that there has been no ‘leaching’ of uranium over time, which could make the dated sample seem erroneously old.

The first principle is relatively straightforward to demonstrate in principle, especially when dealing with complex figurative artworks (for example, naturalistic paintings of animals), which are obviously cultural in origin. With regards to the second, it is not always a simple matter to demonstrate that a calcium carbonate sample of minuscule size collected for U-series dating completely overlies a rock art pigment layer or portion of, for example, an engraving. This is particularly the case when calcium carbonate powder samples are collected in situ using either a scalpel or a hand drill, and where researchers are prohibited from making direct physical contact with, and/or fully exposing, the pigment layer for closer inspection. This method usually results in a cone-shaped sample wherein only the apex or ‘pointy’ portion could be related to the underlying art. As a result, some incorporated calcium carbonate powder could predate the motif. It is sometimes also not clear whether the underlying surface assumed by the analyst to be the rock art has in fact been identified correctly as such—in some instances, it may be a portion of the underlying rock surface, rather than the adjacent artwork itself, that has been dated. Concerning the third principle, the purity of the calcium carbonate sample can be easily and routinely assessed by measuring the ²³⁰Th/²³²Th activity ratios (see below). Finally, the presence or absence of closed-system conditions can be determined by measuring at least three subsamples; to demonstrate a closed system, these should either be of the same age within error or should get progressively younger towards the surface of the sample (see below). In this study, all of the measured ROIs fall within these categories.

Here we report an LA-U-series approach to dating rock art. This method enables us to more readily demonstrate the unambiguous relationship between the calcium carbonate material used for dating and the rock art pigment layer(s) to which it corresponds. Furthermore, the mapping of ²³⁰Th/²³²Th activity ratios across a sample cross-section enables the more accurate identification and selection of calcium carbonate material of the purest quality (that is, the part of the sample with the least evidence for contamination from detrital material), resulting in minor or insignificant age corrections. The latter outcome is impossible to achieve with solution-based methods, as these are based on a much less precise sampling approach that will typically incorporate both pure and dirty portions of a given mass of calcium carbonate. Uniquely, the mapping of U-series isotopes also enables the visualization of key areas or ‘zones’ within the sample in which calcium carbonate material has undergone mobilization of uranium or thorium. Once identified, these problem areas can then be deliberately avoided when selecting ROIs for dating. This is the main advantage of LA-U-series imaging compared with other U-series methods. It also enables the measurement of ROIs that are in clear chronological order, demonstrating closed-system conditions (see below). As has been previously indicated, the other main advantage of this approach to U-series dating is the ability to date calcium carbonate material that formed much closer to the rock art pigment layer. This is desirable, as dating the basal growth layers that accumulated directly on top of the surface of the artwork will potentially increase the minimum age of the associated anthropogenic imagery.

In cross-section, it is possible to visualize zones with discolouration at the surface of samples and sometimes within samples. These zones almost systematically correspond to a pronounced increase in the distribution of ²³²Th/²³⁸U. This ²³²Th/²³⁸U distribution is the easiest to visualize, but there is also usually a corresponding effect on other isotopes. For example, sample BSP4.2 displays a distinct brownish/blackish colour within an area of the sample (Extended Data Fig. 4). This zone corresponds to a marked increase in ²³²Th/²³⁸U and ²³²Th and a small increase in ²³⁰Th/²³⁸U (Extended Data Fig. 4 and Supplementary Fig. 5). This is attributed to the incorporation of detrital material. This incorporation of detrital material possibly occurred inside a porous area that could have been the subject of preferential uranium leaching. To illustrate this, an ROI (ROI-d) corresponding to this area was selected and an age of 38.4 ± 3.6 ka was obtained, thereby showing the effect of identified diagenesis on calculated ages (Extended Data Fig. 4 and Extended Data Table 1). This diagenetic zone is at least 7.2 ka older than ROI 1, corresponding to the layers immediately above the pigment layer and is identical to the solution age obtained for the same sample. The effect of diagenesis can also be seen clearly on LK2 sample sections. Visually, it is possible to identify a porous area in the middle of the sample (Extended Data Fig. 5). This area does not show a brownish-black colour as in sample BSP4.2, but corresponds to a significant increase in ²³²Th/²³⁸U, ²³⁰Th/²³⁸U and ²³²Th, as well as a significant decrease in ²³⁸U (Extended Data Fig. 5 and Supplementary Fig. 3). This porous area clearly underwent significant uranium loss and the incorporation of detrital material. To illustrate this, a new ROI (ROI-d) corresponding to this area was selected and an age of 28.5 ± 3.6 ka was obtained. This is at least 5.3 ka older in minimum age than the other ROIs for this sample. Again, this shows the effect of identified diagenesis on calculated ages (Extended Data Fig. 5 and Extended Data Table 1) and the efficacy of the LA-U-series imaging approach.

Data measurement

A small segment (about 25–150 mm²) of each coralloid speleothem (n = 8) was removed from the rock art panels at Leang Bulu’ Sipong 4 (n = 4) and Leang Karampuang (n = 4) using a battery-operated rotary tool equipped with a diamond saw blade. Each speleothem sample was sawn in situ so as to produce a continuous microstratigraphic profile that extends from the outer surface of the speleothem through the pigment layer and into the underlying rock face. All of the sampled speleothems comprised multiple layers of visibly dense and non-porous calcite in clear association with painted motifs. In the laboratory, the remaining part of the samples from Leang Bulu’ Sipong 4 that were not microexcavated, and the new samples from Leang Karampuang, were sectioned along the growth axis with a diamond saw blade. The samples were then mounted in epoxy resin and polished at a 5 μm smoothness, exposing the pigment layers sandwiched between speleothem layers and/or bedrock.

The coralloid speleothem samples collected in this study formed by accumulation of thin films of water on cave surfaces over a long period of time. When precipitated from saturated solutions, and under ideal conditions, 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, a process in which not all isotopes in the series are decaying at the same rate. Subsequently, ²³⁸U and ²³⁴U decay to ²³⁰Th until secular equilibrium is reached. As the decay rates are known, the precise measurement of these isotopes enables the calculation of the age of the carbonate formation.

All dating work was undertaken at the BIOMICS laboratory in the Geoarchaeology and Archaeometry Research Group (GARG) of Southern Cross University. U-series measurements were obtained using the ESI NW193 ArF excimer laser-ablation unit coupled to a MC-ICP-MS Thermo Fisher Scientific Neptune XT. Each sample was measured by a succession of parallel rasters across the exposed polished cross-section, enabling us to reconstruct an isotopic map of the precipitated calcite. Rasters had a different length to adapt to the irregular shape of the sample using the following parameters for mapping: a square spot size of 44 μm × 44 μm using the infinite aperture of the laser system matched by a translation speed of 21 μm s⁻¹ and integration time of 2.097 s on the MC-ICP-MS Neptune XT. This combination of parameters enabled us to obtain a pixel within <0.1% deformation equivalent to a 44 μm × 44 μm datapoint (the exact translation speed to obtain an exact data-pixel of 44 μm × 44 μm would be 20.982 μm s⁻¹). Other parameters for the data acquisition were as follows: 900 ml min⁻¹ UHP He and 6 ml min⁻¹ UHP nitrogen for the gas flow from the chamber to the ICP-MSs, frequency of 100 Hz for the laser frequency and an average of 1.74 J cm⁻² sample fluence. ²³⁴U and ²³⁰Th were measured simultaneously, with uranium in the centre Faraday cup coupled with a secondary electron multiplier (SEM) and thorium on the L3 Faraday cup coupled to an ion counter (IC). All other Faraday cups were associated with a high-gain 10¹¹ Ω amplifiers (the cup configuration was as follows: L3/IC(230); L2(232); L1(233); C/SEM(234); H1(235); H2(236); H3(238)). Baseline and drifts were corrected using NIST 610 and NIST 612 glass standards, while two corals (the MIS7 Faviid and MIS5 Porites corals from the Southern Cook Islands) were used to correct ²³⁴U/²³⁸U and ²³⁰Th/²³⁸U ratios and assess the accuracy of measurements (Supplementary Information). More information on data measurement is provided in the Supplementary Information.

Image and data processing

Isotopic maps generated using LA-MC-ICP-MS data were produced using the Iolite software package. Data were accumulated in a single file on the MC-ICP-MS Neptune XT system as follows: 5 min background, NIST610 (3×), NIST612 (3×), STD1 (3×), STD2 (3×), STD3 (3×), sample rasters (n×), STD3 (3×), STD2 (3×), STD1 (3×), NIST612 (3×), NIST610 (3×), 5 min background. For sample imaging sequences longer than 2 h, a set of standards (for example, STD1 (3×), STD2 (3×), STD3 (3×)) was incorporated in the middle of the measurement. Data reduction was performed using NISTs to assess drift and the 5 min background on each side of the measurements for baseline. One standard was used for correction of the isotopic ratios, while the other two were used as known values to check data accuracy (including for the matrix effect). Images were produced using a spectrum gradient colour distribution, with either a linear or logarithmic scale (specified for each sample on the isotopic maps). ROIs were carefully selected on the ²³²Th/²³⁸U, ²³⁰Th/²³⁸U isotopic ratio maps and U ppm maps to be as close to the pigment layer as possible, while avoiding diagenetic zones. Data errors were extracted and reported at 2 s.e. ROIs located immediately above the pigment layers were selected to calculate minimum ages relating to the underlying artworks. U-series data were integrated for individual ROIs, resulting in U-series ages and associated errors. Sufficient datapoints were also selected to minimize errors. The integration area of each ROI is reported in μm² (Extended Data Table 1).

It is not uncommon for secondary calcium carbonate to be contaminated by detrital materials, such as wind-blown or waterborne sediments—a process that can lead to U-series ages that are erroneously older than the true age of the sample. This is due to pre-existing ²³⁰Th present in the detrital components. As the detrital/initial ²³⁰Th cannot be physically separated from the radiogenic ²³⁰Th for measurement, its contribution to the calculated ²³⁰Th age of the sample is often corrected for using an assumed ²³⁰Th/²³²Th activity ratio in the detrital component. Given that the detrital component within a cave is often composed of wind-blown or waterborne sediments that chemically approach the average continental crust, the mean bulk-Earth or upper continental crustal value of ²³²Th/²³⁸U = 3.8, corresponding to an ²³⁰Th/²³²Th activity ratio of 0.8—with an arbitrarily assigned uncertainty of 50%—has commonly been assumed for detrital/initial ²³⁰Th corrections. In this regard, the degree of detrital contamination may be reflected by the measured ²³⁰Th/²³²Th activity ratio in a sample, with a higher value (such as >20) indicating a relatively small or insignificant effect on the calculated age and a lower value (<20) indicating that the correction on the age will be considerable. As ²³²Th in the sample is largely present in the detrital fraction and plays no part in the decay chain of uranium, the detrital ²³⁰Th in a sample with a measured ²³⁰Th/²³²Th activity ratio of >20 would make up <0.8/20 (about 4.0% of the total ²³⁰Th in the sample), assuming that the mass fraction of ²³²Th from the detrital component is much larger than that from the authigenic component.

Sometimes, the assumed ²³⁰Th/²³²Th activity ratio of 0.8 (±50%) for the detrital component may not cover all situations. If the actual ²³⁰Th/²³²Th activity ratio in the detrital component substantially deviates from this assumed range, the detrital correction scheme may introduce considerable bias, especially to samples with a ²³⁰Th/²³²Th activity ratio of <20. In such situations, the ²³⁰Th/²³²Th activity ratio in the detrital component can be obtained through direct measurement of sediments associated with speleothems, or computed using isochron methods or stratigraphic constraints. In our case, our samples were relatively pure: the ²³⁰Th/²³²Th activity ratios of individual aliquots were extremely high. Corrections for detrital components were therefore calculated assuming the bulk-Earth values.

A conceivable problem with the U-series dating method is that calcium carbonate accretions can behave as an open system for uranium, in which the element can be leached out of the accretions or remobilized. In such instances, the calculated ages will be too old because the dating method relies on the accurate measurement of uranium versus its decay product ²³⁰Th. In this study, this problem was tackled by avoiding porous samples and by measuring three aliquots (ROIs) from every sample. The ages of these subsamples were in chronological order or of similar ages within error, confirming the integrity of the dated coralloids. If uranium had leached out of the samples, a reverse age profile would have been evident (the ages would have gotten older towards the surface). Age calculations were performed using the UThwigl R package and compared to the IsoplotR (v.6.1) values. Ages are reported with standard errors at the 2σ level.