256 ERNST FREESE 



bo more distant from the original mutation than about 0.2 recombina- 

 tion units. This is only a small portion of the rll region, which extends 

 over about 8 recombination units. One could reduce this possible distance 

 even further, by examining the frecjuency of the original rll mutant, 

 among all r mutants, before and after the cross (by spot tests against 

 the ileletion set) ; this has not been reported so far. 



In another important analysis Benzer (1961) has isolated from a hot 

 spot mutant a few revertants with standard phenotypc and determined 

 a new forward mutational spectrum for each of them. If a revertant 

 arose by a suppressor mutation it should give a different mutational 

 spectrum, either leaving out the original hot spot or else, if the sup- 

 pressor mutation itself is highly revcrtible, giving only the same hot spot 

 and no other ones. In contrast about the same mutational spectrum 

 was found as originally present. 



Even this elaborate analysis does not uneciuivocally prove the 

 existence of back mutations since it is possible, though unlikely, that hot 

 spots correspond to a group of nucleotide pairs. As a further proof it 

 seems desirable that a correlation be made between revertants and the 

 amino acid sequence in the corresponding protein. Interestingly enough 

 even that would not enable a complete proof of genuine back mutation 

 since it is possible that the same amino acid can be coded by two differ- 

 ent nucleotide combinations. Hence the ultimate proof can only come 

 from a partial sequence analysis of the DNA around the mutational site. 



In spite of the above complications several observations on the 

 induction of reverse mutations are clear cut and demand an explanation. 

 When a mutagen efficiently increases the frequency of revertants for 

 certain mutants while it leaves other mutants untouched, it seems 

 probable that the non-inducible mutants contain a base pair change 

 which cannot be reversed by the particular mutagen. The inducible 

 mutants can be further examined in order to decide which of the induced 

 revertants do or do not seem to be caused by a suppressor mutation. If 

 there are induced revertants, which by functional and recombination 

 tests give no indication of a suppressor mutation, it seems reasonable 

 to assume they arose by back mutations. Hence one can conclude that 

 the particular base pair change of these mutants can be induced to 

 back mutate by the mutagen. It does not matter, for this argument, 

 whether none, few, or even many suppressor mutations are induced at 

 the same time. 



2. Transitions versus Other Changef; 



The specificity of the induction of reverse mutations shows up most 

 effectively for the difference between transition and non-transition mu- 



