PROBLEMS OF MEASUREMENT OF MUTATION RATES 61 



this possibility is excluded, because a mosaic would not be scored 

 as a recessive lethal. 



Lederberg: You are also saying that it is not a duplex model, then. 

 You have to make more drastic revisions than this. 



Auerbach: That's the difficulty. But I'm not so sure with these 

 lethals whether there can be some effect on both strands. 



Novick: It's too bad that these experiments can't be done with 

 phage because it would be nice to know how wide these mutants are, 

 what is the nature of such mutations. 



Lederberg: Experiments indicating one strand mutagenic effects have 

 been done with phage (30). 



Novick: I meant experiments of the kind where — 



Atwood: — where you let them sit around and mutate. 



Freese: I would say one has two extreme possibilities. If one uses a 

 base analog as mutagen, it seems rather clear that the increase of the 

 frequency is proportional to the number of cell divisions. In contrast, 

 if one uses nitrous acid, the frequency of mutants is proportional to the 

 ti7ne of treatment, provided that one lets the cells go through one or 

 two more generations that express the change. 



The spontaneous effects may be intermediate; we may have two 

 spontaneous mutagenic effects, one which is proportional to the 

 division time, and another which is proportional to the time the cell 

 stands around. How can we express this quantitatively? We probably 

 have to use two dimensions for the abscissas, one dimension being the 

 division time, and the other dimension being the time of exposure of 

 the cell. 



Zamenhof : Suppose you don't have the cell, but you have mutation 

 of DNA in vitro induced by nitrous acid; would you call it mutation 

 without cell division, if there is no cell? 



Freese: Then you talk about the DNA division. 



Auerbach: Did you ever get mutations in resting bacteria? You 

 found the mutation frequency was proportional to time, but did you go 

 down to complete resting bacteria? 



Novick: What we found was, that as we reduced the gro\vth rate 

 and increased the generation time, the mutation rate remained con- 

 stant per hour until a generation time of about 15 hours. Further 

 increase in the generation time was accompanied by a decrease, a 

 proportional decrease, in the mutation rate per hour. I can understand 

 this fall at very long generation times by imagining that the bacteria 

 cannot regulate themselves to grow at an arbitrarily slow rate, and 

 when the chemostat is operated at generation times longer than 15 



