Rock art is an indicator of occupation of land, and the presence of intelligence and culture within human societies. It offers us a cognitive insight into early modern humans where archaeology is often obstructed from the psychological. Dating rock art, therefore, is of incredibly high importance so that we are able to attribute the works and depicted thoughts and meanings (should there be any) to the correct peoples. Primary methods of dating are essentially impossible due to the nature of rock art, as any piece could have been made with materials that are already of significant age at the time of use, for example charcoal. This means that dating rock art is limited to secondary methods looking at lichens, varnishes, stalagmite formations, and quartz deposits. Using Bednarik’s ‘The Dating of Rock Art: a Critique’ (2002), I have selected a series of dating methods that I believe to be most relevant in answering the question of what difficulties we face in scientifically dating rock art.
The first of these methods is lichenometry. Lichenometry is a method of dating rock art by measuring the growth of lichens and comparing the recorded size to an expected growth rate model to find a date for when growth would have begun. This provides a minimum age for the art as we do not know how long the surface remained untouched by lichens before growth began. Bettinger and Oglesby (1985) put this dating method to test with a series of pit-houses in California. By measuring the thalli size of the lichen on the exterior of the walls they were able to date the construction of the houses, and the interior dated the eventual abandonment, as lichen were prevented from growing indoors by the smoke and dust caused by the inhabitants. Bednarik (2004), supports this dating method explaining that at sites like Foz Côa, if the stones had not been scraped clean, thalli formed over engravings could have been used to obtain minimum dates, and thalli broken by engravings could indicate maximum dates.
However, there are many drawbacks to lichenometry which requires a lengthy and often impractical calibration process, leaving it open to error. Lichen growth is affected by many different ecological factors (Benedict, 2009), some of which are possible to measure, for example exposure to wind abrasion, by looking at the orientation of the site. Others, such as moisture or temperature, are highly difficult to measure as moisture levels can change not only from rainfall but also from fog and temperature levels can fluctuate erratically depending upon the activities of nearby inhabitants. Attempting to factor in these conditions when the site could be anywhere over 1000 years in age is ambitious in the least. Although some younger sites have been calibrated through use of gravestones (Dorren, 2007), which provide their own inscribed date and inherently cancel out the majority of these limitations.
figure 1: Bettinger and Oglesby (1985) part of a table showing dates for interior and exterior walls using lichenometry
Another method of dating rock art is through analysis of the varnishes formed atop them. The varnishes contain atmospheric carbon deposited at the time of formation, and we can also recognise a varnish type by its contemporary environmental conditions (di Lernia, 2010), and so a sample of this can be treated (using the ABA method) to yield a material that will provide a viable date when run through an AMS. This practice assumes, however, that the carbon in the varnish is only atmospheric and that no other sources contribute towards its accumulation (Watchman, 2000). This may be remiss as many things can cause a carbon increase in a small area, particularly if humans are occupying a site after the art has been created for example general respiration and fire lighting. The method also only produces a minimum age as we have no indication as to how long it takes before a varnish begins to form. However, Dorn (1989a) tested the method against pre-dated volcanic sites and it appears that it does produce highly accurate results although there is a fair margin of uncertainty (C14 on charcoal 605+-70 and varnish 710+-150). A serious issue with the methodology that must be considered is that all layers of the varnish can contain carbon and, as we, would obviously hope to take the oldest base layer, incorrect extraction of an earlier layer can seriously affect the results.
Figure 2: Dorn (1989b) image of cut varnish demonstrating different layers that vary in colour due to manganese levels
While both of these methods are possible for use on dating of rock art, neither have really been applied (and published) in any meaningful way in the field yet. OSL, however, has been applied successfully to a site in western Australia. The method requires samples to be taken (generally using coring tubes) and stored without being exposed to sunlight. This in itself is problematic, although possible, and one has to ascertain that the surrounding material was impenetrable to light to ensure that the samples have not been contaminated (Murray, 2007). The sediment must then be sieved to extract the quartz so that it may be cleaned and etched before going through a gamma spectrometer. The method provides an estimation of the time elapsed since the grains were deposited in the stratigraphy by measuring amount of trapped electrons in the chemical structure of the quartz and using the assumed rate of entrapment to find time zero (last exposure to sunlight) (Aubert, 2012). Roberts et al (1997) applied this to wasp nests that obscured three rock art pieces in Kimberley, western Australia. Typically, mud-nesting wasps gather sediments from the surrounding environment to create their nests, which enabled the researchers to dismantle the nests and use the quartz inclusions to date the art. The quartz would have last been exposed to sunlight when the nests were completed and, as testing revealed that no light could reach the centre of the nests, they were perfect for OSL analysis. The analysis provided a minimum age of 100-150 years, although again our understanding is limited by our inability to know the delay between creation of the art and formation of the nests. OSL in itself, however, is a useful and replicable dating method that provides us with a minimum date when carried out correctly. The main drawback of the methodology is that, particularly with rock art, its applicability to a site is rare as quartz grains are often not present and completely hidden from light, or we have no ability to tell whether they were in situ before or after the artwork was created (although one can apply it to buried rock art quite effectively).
The final secondary dating method I will discuss in this paper is U-series dating, which is perhaps the most prevalent in rock art dating despite its specialised needs. The method requires stalagmitic growths to have formed over the art, or for the art to have been drawn atop a layer so that the calcite in the growth can be measured using AMS. Aubert et al (2007) explain this method by telling us that, because flowstone forms from liquid it contains no 230Th (which is insoluble) and so all 230Th found in analysis is the result of the measurable decay of 238U and 234U. In Asturias and Cantabria, Pike et al (2012) applied this method to date the art in palaeolithic cave sites. As a control, all samples had to have an older date on the outer layers than the inner to be accepted as a sample, as flowstones form over time in layers, so this allowed the team to mitigate some effects of contamination. The dates yielded for the El Castillo caves exemplify the importance of dating rock art, the largest date being 40.8kya, making that particular panel one of the earliest known pieces of rock art in Europe. However, there are many limitations associated with using U-series dating to identify the age of rock art. One of these is quite simply a lack of appropriate sites, as flowstone covering art occurs completely by chance, but obviously this cannot be helped. The main issue faced by U-series is contamination: in a study by Taçon et al (2012) “large samples” were collected for both U-series and radiocarbon; the U-series concentration showed unusually low levels of uranium and high Thorium:Uranium ratios which suggested a high probability of contamination and left their results with an extreme margin of error. The final result of this dating method was an age between 15400 and 3400 years ago, demonstrating how contamination can render the method largely useless.
Some primary dating methods have been applied although with limited ability, again due to a lack of appropriate sites. Of course, we are unable to date charcoal usage as the charcoal used could have been over 1000 years old when the artist collected it, however some pigments have been found to contain either human or animal blood. In Australia, two such sites have been dated using blood called Judds Cavern and Laurie Creek (Loy et al, 1990). In this, some pigment was scraped away and measured using radiocarbon AMS. This would yield a very specific and useful date but it has two clear drawbacks: the method is destructive to the art, and any sample is going to be highly susceptible to contamination – especially as one would aim to take as little as possible to avoid unneeded damage. Since then, this date has been heavily refuted by many archaeologists who have been unable to replicate the results, even with samples from the original locations. This has been put down to contamination and the natural inclusion of \”natural calcium oxalate” (David et al, 2013) which was not considered when the tests were originally carried out. This demonstrates the difficulties presented by dating methods as any number of factors that one may not have considered yet can change the result of the tests significantly. A lot of work will have to be carried out to develop this idea into a usable and reliable dating method in the future.
In conclusion, there are many issues faced in dating rock art, the most prevalent of these being an inability to directly date the art without damaging it and to yield dates with confidence that they specifically relate to the time of drawing rather than time of material formation. Lichenometry and varnish AMS present some interesting ways to develop the field that are more widely applicable than U-Series and OSL but have the same limitations in terms of only providing minimum dates and an additional calibration process that can make them incredibly impractical for prehistoric sites. U-series, OSL, and blood analysis are not only limited by applicability but also contamination, however, if one were to find the perfect site that allowed for the dating method and showed no contamination there is no reason that these would not yield useful and replicable dates, although further study must be done to calibrate and understand blood dating before it can be used on a large (or any) scale.
2016-12-13-1481634726