According to R. Layton's hypothesis the structure of "rock art" (cave paintings) differs considerably in terms of frequency of the motives and their distribution between the sites, determined by the religion and the social organisation of the human groups. Such structure is examined again in this collective work through a new perspective based on new documentation from Africa and Australia and on a deeper analysis of the paleolithic parietal art of Western Europe. The conclusion of this study is that the heterogeneity of the sites (geographic distribution of the motives) and the diversity of the iconography (frequency of the motives) are criteria certainly related to the social structure, the believes and the economy of the societies, but they don't allow to decide in favour of a totemic or shamanic organisation.