Younger, 1651). The serial accumulation of pictures in a space had of course been run through a whole range of variations in endless exhibitions, private rooms or even chambers of curiosities. But the crucial media break comes in the 20th century with the mass spread of catalogues of art and other things, and of picture atlases. The basis for this is mass media photographic reproduction, as analysed by Walter Benjamin in his investigations. The work of art in its technical reproducibility takes the form, alongside film, of catalogue production, which becomes a key medium in a new history of pictures and art, located now not in spaces devoted to art, but in lecture theatres or private studios and homes. These media conditions led to Aby Warburg's famous images constellations in his «Mnemosyne Atlas» («Atlas of Memory»), which uses the photographic catalogue in the same way as André Malraux was to do later for his «musée imaginaire,» which he researched from 1935 and published in 1947 as a book with blackand- white illustrations, and Marcel Duchamp as well for his «Boîte-en-valise» (1942). Before this, constructing meaning in alternative displays and algorithmic transformations of image into text and vice

versa had been examined, but Warburg stresses the contextual constellation of history as a visual process presented in three dimensions. Here Warburg was trying «to fuse the systemic ordering function of a typology, the historical ordering function of a type history and the geographical ordering function of a ‹Mediterranean basin event› in a single tableau.« [35] Warburg was concerned with problems of sorting, arranging and displaying relations that we are familiar with today in the concept of a multimedia user interface. The aim was to use specific and complex constellations of photographic reproductions to display relations that were different in each case in such a way that the hidden structures and connections were identifiable visually without textual explanation. So Warburg's picture atlases can also be read as data and relations revealing quite new structures beyond visual or historical-textual evidence, clarifying the media-historical prerequisites of images, for example. [36]

Claus Pias highlighted Warburg's difficulties with representing different relations graphically within an order and on a tableau as «non-contradictory evidence.» [37] But beyond these immanent