MUTAGENESIS 153 



mutants, in the first place? You don't have to go this far with rever- 

 sions. Just measure your reversion rate at the different environments. 



Benzer: Of the same mutagen? 



Lederberg: I'm going to have to clarify this. This was one of the 

 hot spots you talked about before. Now, when you look at these from 

 the point of view of their reversion rate, you find there are three 

 classes of mutants. There is the class that had a rate of 1000, there 

 is the class that had a rate of 1, and there is the class that had the 

 rate of .03. I would say that we actually split this up into three 

 categories. But it is not a very sensitive test. The test has to do with 

 the level of the reversion rate, but could we further split these up by 

 now applying chemicals? 



Benzer: I see what you mean. What appears to be one class could 

 be split into different groups. 



Lederberg: Right, by the use of a battery of mutagenic agents. If, 

 in fact, they cannot, I think that it would be very strong evidence that 

 they are really uniquely just three states, and it would be the 

 strongest support I could think of from this kind of datum that we 

 are really talking about a — 



Neel: How would partial reversions within this cistron fit into your 

 argument. Josh? 



Lederberg: I would say a partial reversion, if it is a reversion, could 

 not take place when you already know the rates; that is, you should 

 be able to list four alternative bases at a given site. In some cases, 

 three of the substitutions will have the mutant phenotype and the 

 original will be the wild type. Then you will have three reversion rates 

 and the wild type. In the cases that you referred to, where you have a 

 partial return to the wild phenotype, it may be that the substitution of 

 one base docs not give you the mutant phenotype but some inter- 

 mediate phenotype. But adding it all together, you should not find 

 more than four states at any one site. 



Auerbach: But, if I understood him rightly, Seymour thinks that the 

 hot spots of his mutagen are influenced by a position effect of the 

 neighboring nucleotides. Would that not be influenced by an exchange 

 in the relevant base pairs so you might expect a splitting, simply be- 

 cause the AT interacts differently with a neighboring GC? 



Lederberg: That is why we can't predict what these numbers will be 

 in advance, because we don't know what the neighbors are and what 

 effect they have on the spontaneous reversion. 



Auerbach: No, I'm sorry. May I show you what I mean? If you 

 have a neighbor here and a neighbor here, if your original neighbors 



