Rosenfeld, in considering the weathering of rock paintings (1985 : 49), states that minerals used in the paints are relatively stable ‘under normal atmospheric conditions, the iron oxides in ochre pigments do not normally change colour or fade’. This is incorrect. Not only can paint residues change their colour entirely (e.g. from white to black, as noted above, or from light-greenish to black), many of the pigments used are chemically mutable under natural conditions. The term ochre does not refer to a specific mineral, it describes a group of earthy mineral oxides and hydroxides of iron, all of which are metastable (Bednarik 1987a). This has been noted elsewhere (‘Patination and petroglyphs’) and it also applies to pigments of iron compounds. Their red, yellow, brown and purple hues are attributable to reflective properties that are determined by the states of hydration, oxidation, reduction, hydrolysis, adsorption and grain size of the various components of such pigments. Therefore iron-rich pigments should be expected to age, striving for equilibrium conditions which can also change with time. Even short-term changes in reflective characteristics, i.e. colour, occur in pictogram paints, and many rock art researchers have observed such changes according to relative air humidity. Such seasonal or even diurnal changes may not be substantial, but they are noticeable. For instance Australian observers have commented on the considerable differences in paint colour prompted by the very humid conditions of the annual wet season. Changes over long periods of time can be much more significant, and it is best not to assume that one can know in what precise colour a paint was originally applied.

These colour changes are not, however, random, there is a trend towards the more stable minerals. Under normal atmospheric conditions, haematite is the most stable of the iron compounds, therefore we need to expect the various components of an ‘ochre’ paint residue to age gradually towards the colour effect of haematite, which is dark red to ‘mulberry’-coloured. It is hardly a coincidence that this is indeed the most common colour found in pictograms, most especially in the hot and dry environments of arid regions which would particularly favour a trend towards the haematite phase.

With the recent advent of digitised colour management systems that are several times as accu-rate as human vision in detecting changes in colour it has at last become possible to monitor these effectively (Mirmehdi et al. in press). This practically super-sedes previous methods of colour monitoring, such as by the use of a chroma meter (Lambert 1995). However, to describe rock art paint residues or pigments as ‘fading’ is of little help in conservation, it needs to be established which specific process is responsible. Essentially three different effects may be involved: the conversion of the pigment involving a colour change, the gradual removal of pigment, or its masking by one or more of several potential agents. Simple unaided visual inspection may not suffice to establish which of these possibilities applies, but without such understanding it may well be futile, if not counterproductive, to design a preservation strat-egy. Pigment conversion, as noted above, can be by any number of processes that alter the chemical constitution of those components of paint residues that determine its colour. Pigment removal, on the other hand, involves a solvent, most likely water or an aqueous solution. Most rock art paints have presumably been applied with water, so unless the pigment component has been well bonded into the substrate’s surface fabric by post-depositional processes, it can readily be remobilised with water. Alternatively a pigment might be quite stable, but its conversion renders it more fugitive, so there may be a combination of two processes at work. Finally, masking of a paint residue occurs generally through the accretion of a more or less translucent mineral precipitate, which perhaps most commonly is a silica skin. It may not affect the pigment directly, it merely conceals it gradually. So more dense the cutaneous deposit, so more transparent, while a porous skin reflects more light and appears whitish. However, if it is sprayed with water to fill the interstitial spaces accessible to it, visibility of the concealed pigment layer tends to improve considerably.

One of the most feared forms of pictogram deterioration is the introduction of cryptogamic growths (see above) into deep limestone caves. This has occurred in European caves with Pleistocene rock paintings after they were exposed to large-scale public visitation over a number of years. Two factors are principally responsible, the introduction of organic matter on the clothing of visitors (pollen, spores etc.) which provided food for bacteria and fungi, and the installation of electric lighting which offered some algae the opportunity to colonise these subterranean environments. Ironically the lighting had been installed to eliminate the use of carbide lamps and torches, precisely to improve preservation conditions for the rock art.


REFERENCES Bibliography of Rock Art Conservation