General Discussion 341 



surely have picked up a mutant on some plates, if the number of 

 mutants in the liquid culture was increasing. 



Slonimski: In the D-arabinose experiment, if you plate 100 cells 

 in a Petri dish you get about lO-^o g. of cell-mass. The amount of 

 sucrar is of the order of 10"^ g. of mass. The analytical reagent grade 

 of "sodium chloride has at best a purity of 99-99 per cent. Let us 

 suppose that you could get D-arabinose much purer than the best 

 preparation of sodium chloride, a rather remote possibility. The 

 mass of impurities added with the sugar would still be much greater 

 (by some 10,000 times) than the initial mass of cells. These im- 

 purities, as well as those coming from other constituents of the 

 medium, may be used for a slow but definite growth, as was shown 

 by Ryan. Therefore, during this critical period mutation followed 

 by selection may very well occur. 



Dean: Prof. Ryan drew attention to the fact that on sugar media 

 a^ar plates one does get microcolonies by growth on impurities. 

 These microcolonies, he thought, contain sufficient cehs to be 

 certain of containing at least one mutant. We think that the 

 exDcriment of plating during the lag phase rules out this possibility. 

 What we measure is the time taken by the colonies to reach the 

 size which is reached by well trained organisms within two days. 

 If the growth were due to a mutant in the microcolony, the time ot 

 appearance of a positive colony should bear no relation to the time 

 spent previously in the liquid medium— if it is plate mutation It, 

 on the other hand, one is selecting mutants in the liquid culture 

 the number of cells should increase during the lag phase. One should 

 be able to detect an increase in bacterial count, just at the end ot 



the lag phase. , . ^u ^ 



Slonimski: My point is that the discrepancy between the reso- 

 lution power of genetical analysis (by studying 100 cells put on a 

 plate or seeded for testing) and the amount of unknown chemical 

 substances present in the test-tube or the Petri dish is so enormous 

 that one just should be very cautious in equating a chemical 

 substance, chosen a priori, as a causative agent of the genetic 



^ Tdlock: We all agree that the colony size, even without training 

 is quite large. I think this growth is partly due to impurities, but 

 that it is probably also due to slow utilization of the lactose. Ihese 

 original parent strains are cryptic mutants and can only use the 

 lactose extremelv slowly. I don't know about D-arabinose. 



The presence of fully trained cells giving rise to rapidly growing 

 colonies on the plate undoubtedly influences the size of colonies ot 



untrained cells. , , ^ x -, 4. 4.u^« 



The colonies of both types (Lac+ and Lac") do come up together 



