Arguably the most influential weathering factor is moisture, which occurs at rock art sites in various forms: as phreatic reservoirs (exposed aquifers, especially in limestone caves), as moving or stationery surface deposits (lakes, rivers, waterholes), as concealed aquifers, as surface run-off, as atmospheric precipitation, as gravitational water percolating through a rock mass, as capillary moisture rising in a porous rock mass, as condensation surface deposit, bound within the crystal structure of component minerals, as interstitial moisture in equilibrium with relative atmospheric humidity, and in the form of air humidity. The interdependence of many of these forms of moisture (e.g. capillary water requires a source such as an aquifer) is crucial in understanding the hydrological regime of a site, which in turn is essential in any remedial work.

There are several deleterious effects caused by most forms of water occurring at a site, while at the same time the presence of some moisture may be essential for preservation. Water travelling within the rock, be it gravitational or capillary, generally acquires solutes (numerous cations are candidates, but chlorides and carbonates are among the most readily soluble in most cases), especially because the pressure within a rock mass is significantly greater than atmos-pheric pressure (it can be hundreds of times greater). Pressure is one of the most important variables deter-mining solubility, the others being temperature, kinetic turbulence, pH and presence of corrosive compounds. This means that when such vadose water re-emerges on a surface, where its pressure suddenly reverts to atmospheric conditions, most of its solute has to be precipitated. Stalactites, for instance, are formed by this process, as are numerous other forms of deposits found at rock art sites.

Weathering zones on rock are primarily the result of solution, hydration, hydrolysis, ion exchange, chelation, oxidation and reduction, leading to the selective removal of mineral mass. Moisture is thus heavily implicated in rock weathering, but as a cause of weathering zones it is usually most effective as direct precipitation. There are exceptions, however, particularly in the case of sandstones, where granular exfoliation of sheltered areas can greatly exceed the effects of weathering processes on surfaces fully exposed to precipitation. Only modest quantities of moisture are involved here, which are sufficient to gradually dissolve the carbonate or silica cement, and which are assisted by the formation of salts that wedge individual grains loose as they expand.

This leads to the next form of moisture-induced rock deterioration commonly seen at rock art sites, called Salzsprengung or salt-wedging. Typically, a zone of salt deposition (called ‘subflorescence’) develops some 5-15 mm beneath and parallel to the surface of a wall, corresponding to the frequent penetration zone of moisture or the extent of a surface zone that is accessible to periodic flushing or drying. The salt might be gypsum, anhydrite, a chloride, nitrate etc., and it has a tendency to increase its bulk, particularly with repeated wetting and drying cycles. This leads to the formation of a surface layer of rock that first becomes ‘drummy’, then lifts visibly and eventually exfoliates as spalls of a thickness corresponding to the alteration zone of the rock fabric. It is in fact this very process that is often the prime cause of the formation of the sandstone shelters of the former Gondwana facies, which I have studied in southern Africa, India, north-eastern Brazil and northern Australia; in all these regions they are extremely rich in rock paintings. Because the results of laminar exfoliation of this type, rockshelters, are the most favoured site type for the production especially of pictograms, such exfoliation is one of the most destructive processes affecting rock art .

Another process causing shattering of rock is attributable to the formation of ice, or to freeze-and-thaw cycles. Liquid water is virtually incom-pressible, having its greatest density at +4ºC; frozen, it expands by 1/11th of its volume. However, crystallisation of water as such does not cause fracturing of rocks, it can only exacerbate already existing cracking by penetrating fissures. In cracks that are less than a few tenth of a millimetre wide, even temperatures as low as -20ºC will not induce water to relinquish its liquid state (Bednarik 1979: 16). On the other hand if openings exceed a few millimetres, congelifraction is not effective because the expansion of water through crystallisation will merely force water out. Rock fatigue due to freeze-and-thaw cycles is perhaps attributable to either the compression of air sealed off in voids by the developing ice (Tricart 1969) or to regelation (Schmid 1958). Be that as it may, the extremely high content of cryoclastics in stadial, entrance-near deposits in caves shows tellingly how effective congelifraction is on at least some rock types. Porous rocks (limestone and dolomite, for instance) are subjected to severe fracturing (Bednarik 1970), as are schists because of their laminate structure inviting penetration by gravitational water. I have suggested that the principal reason for the absence of Pleistocene cave art in central Europe is that the caves there, in contrast to south-western Europe with its more temperate climate, were subjected to extensive congelifraction during the final Würm stadial (Bednarik 1994a). Indeed, the only Palaeolithic rock art ever found in central Europe was found on exfoliated clasts in German caves, having been destroyed by freezing and/or regelation (Hahn 1990; Conard and Uerpmann 2000).

Condensation water forms in caves and rockshelters when the relative air humidity is high and the ambient air temperature is greater than the rock face temperature. The dew point is the temperature at which air becomes saturated and dew is precipitated. Limestone caves frequently have a relative air humidity of 95% to 99%, yet condensation is not necessarily common in caves, particularly those where the speleoclimate is almost entirely independent of the atmospheric climate. Nor is the water, which condenses in distilled form, capable of dissolving the limestone in the absence of carbon dioxide. In a deep cave system, air temperature is buffered by rock temperature and therefore very stable. There is practically no convective temperature exchange in caves of a high air residence rate, and the temperature differences necessary to induce condensation do not exist. Condensation is more likely to appear near the entrance of a cave, or in rockshelters. It forms as water beads or as a film that can not only cause significant damage to paint residues (paints were often based on water as the solvent), it tends to attract cryptogamic growth within the reach of light (artificial or natural), and it can deposit a variety of dissolved salts. In the case of carbonate rock, such surface water deposits can even form carbonic acid in the presence of carbon dioxide, and convert the rock to water-soluble and mobile bicarbonate.

Surface run-off is another factor frequently implicated in the destruction of rock art. On pictograms it can be highly effective, removing unconsolidated pigments rapidly. Where it occurs regularly, the establishment of mosses, fungi and lichens may be an unfavourable result. Mineral accre-tions, including ferromanganous crusts, may be deposited from solute-rich solutions. Rainwater is never pure water, it acquires numerous ions already before it reaches the ground, many of which assist it in converting or dissolving minerals. There is a distinct tendency for such inclusions, which are usually chlorides, sulphates, nitrates, phosphates and bicarbonates, to effect a lowering of the atmospheric water pH. It is well known that this has been exacerbated by the industrial development of recent centuries. However, organic acids are also present in rainwater. After noting the high contents of formate, acetate and proprionate in cloud and rainwater in tropical Australia (Noller et al. 1987), which is attributable to large-scale vegetation burning, Watchman (1990) suggested that these acids may cause the formation of oxalate crusts. Isoprene, found in many plants, is released into the atmosphere, converted by hydroxyl radicals to formaldehyde, which is then oxidised to formic acid. Similar photochemical reactions can convert other breakdown products of isoprene into oxalic acid, which can convert carbonates, including some rock art pigments, to oxalates.

Mass-exfoliation of discrete rock facies in stratified deposits can be caused by either gravitational or capillary moisture travelling within the rock. It is one of the more pernicious rock art conservation problems, because it not only destabilises otherwise perhaps quite well-preserved panels, but it is particularly difficult to address. Attempts to control capillary water, in particular, have a long history of failure. A comparatively stable rock art panel may be disintegrating because it depends on the structural support of another, less stable facies that is rapidly dissolving through moisture. Two different regimes are known, exemplified by two sites I have investigated. In the first case, at Shishkino in central Siberia, the rock art occurs on a vertical cliff, on a horizontal band of sound sandstone. Immediately below this stratum is a layer of argillaceous sandstone (almost mudstone) which decays rapidly from the action of internal water seepage. As it disintegrates, it leaves the stratum without support, which will eventually collapse from its own weight. In the second case, at Mootwingee in eastern Australia, a steep slope dipping at about 19º consists of relatively resistant siliceous sandstone. It bears the rock art, but is underlain by a layer cemented mainly by gypsum and hydrated aluminium silicates, both of which are removed readily. The upper layer develops fissures as a result, which in turn increases access by water and weathering of the lower stratum. Because of the slope’s steepness individual blocks, which ‘float’ on this increasingly unstable support, then slide downwards. While this is seen as a form of physical weathering, the true source of the deterioration is primarily chemical action of moisture. Similar effects can work at a smaller scale, for instance Ollier’s (1969) ‘dirt cracking’, which is essentially attributable to the swelling of clays deposited in fine rock fissures.


REFERENCES Bibliography of Rock Art Conservation