X. TMV STUDIES IN GENETIC CODING 507 



ena were observed, four showed a single "secondary" amino acid ex- 

 change. This could be due to some heterogeneity in the primary mutant, 

 with a minor component (which then could be attributed to the original 

 modification reaction) being favored by the test conditions. On the other 

 hand, these observations may also signify that some of the chemically 

 evoked mutants are rather unstable and readily undergo additional 

 changes due to spontaneous mutations in the course of replication. The 

 corollary of this hypothesis would be that the spontaneous mutation 

 frequency is much higher for RNA than generally presumed, but that 

 this fact is usually concealed by the lesser viability of the vast majority 

 of the mutants as compared to the parental strain. If this were the case, 

 then the observed mutants showing very different composition from the 

 parental virus might be attributed to a series of spontaneous mutational 

 events within a low- viability population, leading ultimately to a stable 

 or rather a viable virus which then, not too surprisingly, would tend to 

 resemble a known natural strain. In that case no direct relationship 

 between the mutagenic reaction and the final stable strain would be 

 expected. 



These considerations probably do not apply to the strains showing 

 only one or two exchanges, and yet among these also the same amino 

 acid replacements were frequently observed to recur, regardless of the 

 mutagenic reagent used. 



In conclusion, then, no evident correlation was detected between the 

 presence and extent of amino acid exchanges and either the nature of 

 the modifying reaction or its intensity. The only clear correlation which 

 was noted in this work was that amino acid changes were detected in 

 almost all mutants giving local lesions on A^ sylvestris, and in only one 

 mutant not showing this distinct biological property. It is this recog- 

 nition which enabled us to select more promising mutants for chemical 

 study, and thus to obtain relatively more data on amino acid exchanges 

 than could be obtained by random analysis of mutants irrespective of 

 the nature of their symptom alterations. 



Parallel to our studies Wittmann analyzed a considerable number 

 of nitrous acid mutants (1960a, 1961). His techniques differ from our 

 in the following regards. He modified the virus directly, rather than the 

 RNA, and has used only nitrous acid as the mutagen. As starting mate- 

 rial he frequently employed a strain (A-14), rather than common TINIV. 

 For the detection of mutants, Wittmann relied on tlie transfer of single 

 primary lesions from Xanthi to N. tabaccum (symtom mutants) avoid- 

 ing selection on A'', sylvestris. The reason for this was his desire to 

 minimize the danger of secondary mutations in the course of repeated 

 transfer. After isolation of the progeny protein, Wittmann directly 



