Beijerinck's entirely new concept, launched in 1898, of a filterable contagium vivum fluidum which multiplied in close association with the host's metabolism and was distributed in phloem vessels together with plant nutrients, did not match the then prevailing bacteriological germ theory. At the time, tools and concepts to handle such a new kind of agent (the viruses) were non-existent. Beijerinck's novel idea, therefore, did not revolutionize biological science or immediately alter human understanding of contagious diseases. That is how bacteriological dogma persisted, as voiced by Loeffler and Frosch when showing the filterability of an animal virus (1898), and especially by Ivanovsky who had already in 1892 detected filterability of the agent of tobacco mosaic but kept looking for a microbe and finally (1903) claimed its multiplication in an artificial medium. The dogma was also strongly advocated by Roux in 1903 when writing the first review on viruses, which he named 'so-called "invisible" microbes', unwittingly including the agent of bovine pleuropneumonia, only much later proved to be caused by a mycoplasma. In 1904, Baur was the first to advocate strongly the chemical view of viruses. But uncertainty about the true nature of viruses, with their similarities to enzymes and genes, continued until the 1930s when at long last tobacco mosaic virus particles were isolated as an enzyme-like protein (1935), soon to be better characterized as a nucleoprotein (1937). Physicochemical virus studies were a key element in triggering molecular biology which was to provide further means to reveal the true nature of viruses 'at the threshold of life'. Beijerinck's 1898 vision was not appreciated or verified during his lifetime. But Beijerinck already had a clear notion of the mechanism behind the phenomena he observed. Developments in virology and molecular biology since 1935 indicate how close Beijerinck (and even Mayer, Beijerinck's predecessor in research on tobacco mosaic) had been to the mark. The history of research on tobacco mosaic and the commitments of Mayer, Beijerinck and others demonstrate that progress in science is not only a matter of mere technology but of philosophy as well. Raemaekers' Mayer cartoon, inspired by Beijerinck, artistically represents the crucial question about the reliability of our images of reality, and about the scope of our technological interference with nature.