DELAYED EXPRESSION OF MUTATIONS 459 



these passed through 25 and 15 generations respectively (the relevant data 

 are given in table 5). The mutation rates calculated by methods 1 and 2 



Table 5 



Generations between resting stage and the approximate time of the first mutation in series of liquid 

 cidtures started from widely different inocula showing that mutation rate is independent of this vari- 

 able over the range of 6 to 18 generations. {Data from table 1.) 



showed no effect due to the difference in number of generations from resting 

 stage. Thus the time of the supposed transitions would have to be a function 

 of the approach to saturation, and not of the number of generations from rest- 

 ing stage. This means that mutation rate would have to be of the order of 

 3 XlO -8 during the whole of the logarithmic growth phase, dropping to some- 

 thing like .4 X10 -8 during the last few divisions. 



It will be shown later that there are apparent variations in mutation rate 

 during early growth, but that they are associated with the first few divisions 

 after resting stage and occur too early to have any bearing on the immediate 

 problem. 



It is thus evident that the possible effect of a low mutation rate during the 

 later part of growth can be eliminated by confining one's tests to the period of 

 rapid growth; and critical evidence for or against a possible delay can be ob- 

 tained from estimates of mutation rate based on the number of resistant clones 

 appearing during rapid growth. 



EVIDENCE FOR A DELAY BETWEEN MUTATION AND PHENOTYPIC 

 EXPRESSION (METHOD 4) 



It now remains to determine the rate of appearance of resistant clones dur- 

 ing rapid growth, when no approach to saturation is involved. 



If the rate of appearance of resistant clones is high (3 X10 -8 ), the discrep- 

 ancy between methods 1 and 2 can be interpreted without assuming a delay 

 between mutation and phenotypic expression. If, on the other hand, rate of 

 appearance of resistant clones is low (.4X10~ 8 ), then from the previous evi- 

 dence the number of individuals in an average resistant clone must be greater 

 than expected on the basis of the number of generations passed through after 



6i 



