1974 we developed a method for the assembly of active 50S subunits from its 35 components (total reconstitution, 24). With this method and a newly developed procedure for analyzing assembly cooperativity we identified the two ribosomal proteins, which are able to initiate together the assembly process of the large subunit (1982; 25). Similarly, two assembly initiating proteins of the small subunit could be identified (1988; 26). We determined reconstitution intermediates and the rate-limiting steps of assembly (1976, 1978; 27,28), discovered the "assembly gradient" in vitro (1977; 29), which describes the necessity to couple transcription of rRNA and ribosome assembly, and confirmed it in vivo (1993; 30). The assembly gradient is the only explanation hitherto, why ribosomal assembly has to occur – without any exception – in the nucleoli of eukaryotic cells. Eventually we identified early (1978; 31) and late assembly proteins (1990; 32), which unraveled the network of detailed assembly dependencies (assembly map, 1986; 33).