86 MUTATIONS 



for forward mutations by the replication method, and he did ob- 

 tain mutations. While Vielmetter in phage got a one-hit curve for 

 mutation, which is all right on his hypothesis that lethal hits are very 

 frequent, Kaudewitz got a two-hit curve. This agrees with the two- 

 strandedness of DNA, but then he should have had mosaics and he 

 had none. Altogether he had 539 mutants at survival rates ranging 

 from 10 - to 10"^. None of these was a mosaic. Then he thought of the 

 objection that Dr. Benzer raised just now, namely that he might have 

 missed small non-mutant sectors. So he took 20 mutants that had been 

 obtained with low doses, where lethal hits on one strand accompanying 

 a mutation are least likely, and tested them fully by making suspen- 

 sions and plating them. None of them was a mosaic. Actually, this re- 

 sult would need to be corrected for spontaneous mutants, which may be 

 nonmosaic; for one of his 20 fully tested mutants had arisen without 

 treatment, and on the basis of the frequency of mutations in the un- 

 treated series, one would expect that at least 6 of the others were also 

 of spontaneous origin; so this would leave only about 13. Still, it is 

 puzzling that none of these was mosaic. I should add that there is 

 nothing in the method that would obscure the occurrence of mosaics; 

 for with the same method P^- had been found to produce many mosaics 

 in experiments likewise carried out by the Tubingen group. In those 

 experiments, about 30 per cent mosaics had been obtained at a survival 

 of 10-^. 



Altogether then, the experiments on bacteria are difficult to interpret. 

 The two-hit curves for mutation and killing suggest that the bacteria 

 were binucleate. Kaudewitz says they were not, but I cannot see how 

 he can feel so sure about it. He draws the conclusion from observations 

 by Witkin and from his own experiments with P^-, but in all these 

 experiments there had been different conditions of culturing and treat- 

 ment. However, even if the bacteria in the nitrous acid experiments 

 were binucleate, where were the expected mosaics? In any case, these 

 data cannot be taken as showing that nitrous acid in bacteria acts in 

 the same way as in phage, namely, by deaminating a base in one of the 

 two strands of DNA. 



Zamenhof: In our paper (84) on the induction of mutation in bac- 

 teria by nitrous acid, we also did not find mosaics. But I want to refer 

 here to the work on heating (85), where bacterial spores, immediately 

 after heating and without an intermediate cultivation, were plated, and 

 one obvious way one could detect auxotrophs in this case would be if 

 the second strand in DNA and the second or third and fourth nuclei 

 were all destroyed. Up to 10 per cent of auxotrophs were discovered 



