Establishing the age of rock art is usually considerably more complicated than dating materials from archaeological layers. These complications arise because of the difficulty in confirming the association between the artefact (in this case, the art) and the material being dated. For a well-sealed archaeological layer it is often accepted that dates from one or two bones, or fragments of charcoal, or better still a sequence of dates above and below the relevant deposit provide an age or age-range for the other artefacts in the layer. In exceptional circumstances decorated blocks become detached from the walls and ceilings of caves and rockshelters, and are incorporated into datable archaeological sediments, such as with the Solutrean sculpted panel from Le Roc de Sers (Tymula 2002). In the vast majority of cases, however, engraved or painted rock art is seldom part of a depositional sequence and thus it is extremely rare that the age can be as well constrained. Dates of 'rock varnish', weathering rinds, or calcite deposits that overlie the art can only provide minimum ages, whereas radiocarbon dating of the pigments can only provide a maximum age, that is, the date of the production of the charcoal. Potentially the time delay between the preparation of a pigment (e.g. the growth of a tree and the preparation of the wood to a charcoal-based pigment) can be assumed to be insignificant, but the use of old charcoal cannot be ruled out (e.g. see Pettitt and Bahn 2003). Furthermore, the dating techniques themselves have potential inaccuracies. With the very small quantities of carbon available from rock art, contamination of radiocarbon samples is a big issue. Other methods for example, cation-ratio dating (e.g. Nobbs and Dorn 1988), are affected by factors such as the average temperature, which at best can only be estimated.
Even when it can be assumed that the dating methods are chronometrically accurate, the relationship between the material dated and the art itself must be completely secure. The date of a calcite layer only provides a minimum age for the art if it overlies the art and is not in potential contamination with earlier flowstones. In the case where the art was made on already deposited calcite and subsequently covered by further calcite, it may not be possible to remove samples from only those layers above the art, and thus a date for the calcite will not relate to the date of the art. It may seem a trivial assumption that a black charcoal-based pigment comprising part of a figurative depiction has an obvious temporal relationship to the figure, but the possibility of retouching, overdrawing, or later deposition of soot from lamps may complicate the picture. Attempts to date rock art, therefore, must proceed with caution, using only the securest samples and controlling for contamination. They should certainly be regarded as developmental, and not routine, chemistry.
Inaccuracies in dating may go undetected or may be in obvious conflict with the stylistic interpretation or the archaeology in the immediate vicinity. In the case of conflict it would be a circular argument to rely entirely on the stylistic attribution of a date over absolute dating methods, but nor should we ignore stylistic interpretation simply because we have absolute dating methods.
The rock art at Creswell takes the form of engravings directly into the limestone bed-rock, and there is a lack of datable pigment that one might obtain from painted panels. Fortunately, a number of the images were overlain by thin veneers of precipitated calcite (flowstone) which is datable by uranium-series (U-series) disequilibrium dating. In April 2004 a number of samples were taken with this aim in mind, which should establish a minimum age for the underlying art, given that the flowstone accumulated some time after the art was produced. Here, we present the results of this project.
Was this article helpful?