The existence of nonsense or stop codons was suggested by the mutant phenotype of certain combinations of (+-) double mutants (see previous session). Benzer had observed that certain rII mutants can behave as wild type in some restrictive strains that for example do not allow the growth of rII deletion and frameshift mutants; he called these mutants ambivalent rII mutants and the strains suppressor strains. In the first paper, he used again the deletion that fuses rIIA and rIIB while preserving rIIB function. Base analog revertible mutations in the N-terminal region of rIIA were combined with the deletion. Some of the mutations, called missense mutations, do not affect the rIIB function of the fused gene. Others, in striking contrast, abolish its rIIB function in restrictive strains but not in the suppressor strains. One of the major conclusions of this work is that the genetic code is under the genetic control of the organism itself. This control is achieved by the genes that encode the tRNAs and the amino acid synthetases of the organism. Brenner discuss the results observed in several bacterial and viral systems that describe nonsense mutations and their bacterial suppressors. They propose a unifying nomenclature for amber and ochre mutations. They show that amber suppressors only suppress amber mutations whereas ochre suppressors, which are usually weak, suppress both amber and ochre mutations. Variable results obtained with different suppressors are explained by the nature of the amino acid inserted by each suppressor as well as by context effects. The major mutagen used in this work is hydroxylamine, which can modify C residues so that they are recognized as T during transcription and replication. Since the rII genes have to be expressed before replication, a modified C will only exert its effect if it is present on the DNA strand that is transcribed. Using hydroxylamine, they show that neither amber nor ochre mutants can be induced to revert with hydroxylamine either immediately or after on growth cycle on a permissive strain. Since ochre can be converted into amber by treatment with 2-aminopurine, a base analog mutagen, the two codons must have two common bases (UAx) and the third differs by a transition (UAG versus UAA). In a brilliant hunt for forward amber and ochre mutations induced by hydroxylamine, they show that the two codons must have both a U and an A. This genetic data is combined with biochemical study of proteins produced either by suppression of amber mutants or by mutagen induced reversion of these mutants. The data fits perfectly the biochemical deciphering of the code that was performed by Nirenberg et al.