One of the most satisfying experiences of phytosociological work is the moment when out of the mass of amorphous data begins to crystallize a lattice of general principles and an understanding of vegetation and its ecology. But if one works still further and gathers more and more knowledge in one and the same field, one reaches a deeper level. The more one knows, the more the boundaries of classes are broken down, the crystals loose their defined shape again and one begins to understand the impossibility of classifying nature in a generally satisfying way. If one tries to retreat to the classes of individuals called species, one has soon to admit that, here too, with growing knowledge the boundaries become insecure, especially where the notoriously heterogeneous weed species are concerned. Here one approaches the boundary between science and philosophy. As this is to be a scientific paper we have to stop here and try to maintain classification as a mere practical means to get a general view and to summarize and pass on our knowledge to others.