DELAYED EXPRESSION OF MUTATIONS 457 



mutants of a similar kind, since these occur much more frequently, and cultures 

 containing mutants and non-mutants in suitable proportions for competitive 

 growth experiments can be obtained without the necessity for first isolating 

 the mutant strain. Evidence has been obtained in this manner in connection 

 with a separate study (Newcombe and Scott) which will be published later. 

 Six independent mutant clones were tested over prolonged periods of growth 

 (18 to 20 generations) in competition with the corresponding non-mutant 

 strain. In some cases the proportion of mutants was found to remain unchanged 

 and in others to have declined slightly, the factors of change ranging from 

 1.20±40 down to 0.16+ .05. In no case did the proportion of mutants increase 

 significantly during growth. 



Thus, provided one assumes that the radiation induced mutants are identi- 

 cal to those occurring spontaneously, evidence from this source is in agreement 

 with the above conclusion that the high estimates of mutation rate from meth- 

 ods 2 and 3 cannot be due to a differential favoring the mutants during the 

 period of rapid growth. 



These experiments do not eliminate, however, the possibility that during 

 the approach to saturation, when the environment has altered considerably, 

 there may be conditions favoring the mutant strain. An increase in the propor- 

 tion of mutants during this period of approximately six times, would be suffi- 

 cient to produce the observed high estimates of mutation rate arrived at from 

 the numbers of individual resistant bacteria. 



In order to test this possibility, an experiment was designed in which the 

 conditions associated with the approach to saturation could act repeatedly 

 on a bacterial population passing through approximately five to six generations, 

 thus accentuating any selection. 



Five replicate test cultures, each containing in the first instance one cc of 

 broth, were incubated over a period of three days, during which time the 

 amount of liquid medium was doubled at regular intervals and finally brought 

 up to fifty cc. If the approach to saturation in the test cultures used in connec- 

 tion with methods 2 and 3 is accompanied by a six-fold increase in the propor- 

 tion of mutants, one would expect from the above treatment an increase of 

 much more than six times, in addition to any mutants which arose from spon- 

 taneous mutation. Also, estimates of mutation rate obtained from these cul- 

 tures using method 3, should be very much greater than the corresponding 

 estimates from normal test cultures. 



Such estimates of mutation rate from the above experiment are presented 

 in table 4. It will be seen that these are not appreciably increased by keeping 

 the cultures under conditions approaching saturation throughout the whole 

 of the growth period. 



In evaluating these data it will be noted that all assays are of numbers of 

 viable bacteria, and that if there is appreciable death due to the conditions of 

 growth, the number of cell generations will be underestimated. This problem 

 is considered in detail in a later section, where it will be shown that the effect 

 of undetected cell death is to increase the estimate of mutation rate obtained, 

 to a value above that of the true rate. Thus the experiment is weighted against 



59 



