Some of the basic problems presented by the rapid evolution of a universal genetic code can be resolved by a mechanism of co-evolution of the code and the amino acids it serves. The genetic code of protein synthesis assigns 64 triplet codons to amino acids and translational signals. Since the coding is universal amongst all prokaryotic and eukaryotic species for which there is evidence (1), it appears to have remained unchanged throughout the course of biological evolution. The code therefore stands as a faithful record of the events which at an early stage of the development of life on this planet had led to its emergence, selection, and fixation. Nirenberg et al. (2) first observed that some of the amino acids coded for by contiguous codons in the code are physically or metabolically related to each other. Sonneborn (3) and Woese (4) suggested that the physical relatedness could have accumulated through natural selection by virtue of the protection it offered against lethal ,mutations or translational errors. As for the question of whether or not the metabolic relatedness might be traceable to the origin of the code, an important step toward its resolution was taken when Pelc (5) and Dillon (6) suggested that the distribution of the codons amongst the amino acids could have been guided by potential conversions between the amino acids. Unfortunately, the postulated conversions were entirely based on the structural relatedness of the amino acids, and contradicted many of the actual metabolic conversions already known to occur in organisms. All of these proposals postulated an evolutionary origin for the genetic code, but did not provide any unambiguous prediction of codon distribution that can be tested against the distribution found in the code. This untestability renders it impossible to define the extent to which the apparent physical or structural relatedness between neighboring amino acids in the code may be the consequence of chance rather than evolution. Given the abundance of such relatedness amongst amino acids, even completely random allocations of code words would not be devoid of all elements of relatedness.The co-evolution theory (7) attachment to tRNAs are well known; and the covalent transformation of an amino acid on the tRNA also finds illustration either enzymically in the formylation of Met-tRNA (8), or nonenzymically in the desulfurationof . Since the biosynthetic relationships between amino acids in present day organisms are clearly defined, predictions could be made on this basis regarding the distribution of the codons for both precursor-product and sibling pairs of amino acids in the code. The excellent fit between both types of predictions and the genetic code excludes chance and establishes evolution as the originator of the code. The necessity for nonevolutionary origins such as frozen accident or infection from an extraterrestrial source (10) is removed.Since evolution as we know it operates within the framework of a universal genetic code, the evolution of the universal cod...