Once made, many San images continued to perform significant functions; they were not made for one occasion only and then ignored. The rock shelters were not simply ‘galleries’ packed with objets d’art. As I have already argued, many paintings were things-in-themselves, not just pictures of things that existed elsewhere, perhaps out in the countryside or around a camp-fire. Again, there is, unfortunately, little ethnographic information on precisely what happened to these potent images after they had been made. Given the three stages that I have so far described, it seems unlikely that paintings would simply have dropped out of the ambit of San ritual and belief. On the contrary, such evidence as we do have suggests that they continued to play an important function.

In the first place, M’s statements suggest that the images were important visually, but in a specific, non-Western way. She demonstrated how, long ago, San people danced in the painted rock shelter to which she took Peter Jolly and me and how they raised their arms and turned to the paintings when they wished to intensify their potency (amandla). As they danced and looked at the paintings, power flowed from the images and entered into them. This was, I believe, her way of saying that the sight of the paintings deepened the dancers’ trance experience. The images were reservoirs of n/om.

M then expanded on this notion that the images themselves contained potency. She said that if a ‘good’ person placed his or her hand on a depiction of an eland, the potency in the painting would flow into the person, thus giving him or her special powers. To demonstrate how this was done, she held my wrist and arranged my fingers so that my entire hand was on a depiction of an eland. As she did so, she unnervingly cautioned that, if a ‘bad’ person did this, his or her hand would adhere to the rock and the person would eventually waste away and die. Her actions recalled the ‘laying on of hands’ during a San curing dance: shamans place their hands on people and draw sickness out of them and into their own bodies; they then expel the sickness through a ‘hole’ in the back of the neck, and it flies back to the spirit world from whence it came. A comparable process of transference of supernatural essence, it seems, could take place between a person and a painted image imbued with power. In the case of paintings, the images were thus mediators between the material world and spiritual realms and entities. The importance of touching, and not merely looking at, rock paintings is supported by evidence from various parts of southern Africa. In the Western Cape Province patches of paint have been rubbed smooth (Yates and Manhire 1991). It is not entirely clear with what the patches were rubbed, but the smoothness of the rock, particularly in the centre of the patches, is easily discerned. In addition, some Western Cape images were rubbed, possibly both before and after the paint dried; their outlines are characteristically blurred and smeared. Rubbed images have also been recorded in the Waterberg in Northern Province (Laue 2000; Peters 2000). Similarly, the making of positive handprints that are common in parts of the Western Cape Province and some other regions was probably closely associated with ritual touching of the rock rather than (exclusively) with the making of ‘pictures’ of hands (Lewis-Williams and Dowson 1999: 108; Lewis-Williams and Blundell 1997).

There is thus evidence that some of the paintings were not made merely to be viewed. After they had been made, they continued to be involved in rituals, and physical contact with some images facilitated the acquisition of potency. M’s remarks suggest that the touching of paintings was not open to everyone and that for some people it could be hazardous. Significantly, she was implying differential access to the art and concomitant social differentiation. More than that, the fixed visions already on the rock face probably contributed, as I have suggested, to the dancers’ hallucinations, informing and constraining the stream of mental images that the human nervous system produces in deep altered states of consciousness. All in all, the painted images did not merely record events and experiences. They became sources of potency and active components in a complex ritual of dancing, singing and clapping that controlled the spiritual, or hallucinatory, experiences of shamans and, possibly other people’s as well. Painted rock shelters were nodes on the landscape where the spirit realm was especially immanent (Ouzman 1996)