. tae _ \ meOe Che ae ee re ne Porp/Baruiie/Hinscu/ConrRaD 216 substitution of isolcucine may be caused both by infrequent nongenctic errors and by somatic mutations. However, isoleucine substitutions for amino acids whose codons require more than one base substitution to change into isoleucine codons can be considered to be due primarily to translational errors because the likelihood of two base substitutions within the same codon is the squarc of the frequency of a base substitution at one nucicotide, which would be very infrequent. Thus, isoleucine substitution for amino acids like glycine or alanine can be considered to result from translational errors only. Some aminoacids are coded by triplets, some of which can and others which cannot mutate from nonisoleucine to isoleucine codons by single base substitutions, c.g. lysine, arginine, threonine and serine. Substitution of isoleucine for such amino acids would be by nongenetic errors at some positions, but by somatic mutations plus nongenetic errors at other positions. The distinction of these should be readily made by treatment of experimental animals with base substitution mutagens; marked increases in isoleucine substitution should occur only at positions where single base changes can mutate nonisolcucine codons into isoleucine codons. By comparing the substitution frecuencies of isoleucine for lysine at positions which are and which are not affected by a base substitution mutagen, it should be possible to access the relative contribution of translational errors and somatic mutations toward the formation of erroneous proteins during protein synthesis in young and old animals, In line with the above arguments, the data of LorrrieLp and VANDERJAGT [7] are pertinent and interesting. If their data are handled in a manner similar to ours, i.c. to calculate the valine substitution for all the other amino acids in the tetra- and octapeptide (except for the C-terminal amino acids © because valine substitutions for lysine would not yield the appropriate tryptic peptides), their »verage substitution frequency would be 7.5 x 10-* per amino acid residue compared with our value of 2.99 x 10-5 per amino acid residue. These valucs are in good agreement and the twofold differences may actually be due to specics differences. LOFTFIELD [6] has found that the majority of the valine incorporated in the tryptic peptides of the rabbit a-chain is located at positions where isofcucine is the coded amino acid (personal commun.). This would seem to support their view that most of these substitutions are caused by translational crrors. However, amino acids that are likely to be involved in translational errors are the same ones which are coded for by triplets wheresingle bascs distinguish nonisoleucine and isoleucine codons. Of greater interest is the fact that the valine substitution in the tetrapeptide was lower than in the octapeptide. The octapeptide contains 3 glycine residues,