444 G. S. WILSON 



generation must therefore have been 1.71. Using the formula of 



lo£C & — lo'^ (I 



n = — -. ^ _° — it is found that in 240 minutes there were 12.2 



log 1 .71 



generations, which gives a generation time of 19.7 minutes. 



Calculated by the old formula n = —^, — the number of 



log 2 



generations is only 9.4, and the generation time is 25.5 minutes. 

 It is clear therefore that the generation time is in general shorter 

 that that hitherto recorded. To give an idea of the probable 

 figure, the average generation times for the whole 16 counts dur- 

 ing the logarithmic phases has been worked out; it comes to 

 21.05 minutes. 



Taking all the facts into consideration, it seems that in cultures 

 of Bad. suipestifer there is a normal death rate, even during the 

 period when the maximum rate of growth is proceeding. The 

 extent of this will vary in individual cultures. In some it is as high 

 as 43 per cent, in others it is only 20 per cent or 10 per cent, 

 while finally in a few it is for a short period actually nil. 



CONCLUSIONS 



1. A method for estimating the number of the total and the 

 viable organisms in a culture medium is described. It is claimed 

 that the experimental error involved in each count probably does 

 not exceed 5 per cent. 



2. Applying this to the study of the relation existing between 

 the living organisms in a culture and the total number of organ- 

 isms alive and dead, it is seen that even during the logarithmic 

 phase the percentage of viable organisms seldom rises above 90 

 per cent of the total. 



3. To explain this a hypothesis is advanced which supposes that 

 in an in vitro culture of Bad. suipesiifer there is a normal death 

 rate amongst the bacteria, even during the logarithmic phase of 

 growth. 



4. Assuming the presence of a normal death rate, it has been 

 possible to calculate the generation time on an altered basis, and 



