POPULATION GENETICS 



sary by a second isolation from a single colony, and then used 

 for the preparation of a new stock culture. It is a rather 

 different situation when the requirement is not to maintain 

 one quality but to preserve all the genetic qualities, including 

 some, such as virulence, which would require very elaborate 

 tests to check at each transfer. Here the logical approach 

 would be to assume that mutation is a rare occurrence and 

 that when the culture population has reached its peak there 

 will be only a small fraction of mutants. The random choice 

 of a single colony for transfer will ensure that the genetic 

 character is maintained for all but qualities which are highly 

 subject to mutation. Such continuously repeated re-isolation 

 for transfer of stock cultures has of course been completely 

 replaced by storage of freeze-dried cultures. 



Re-initiation of a multiplying culture from a single repre- 

 sentative of an old one may seem a very artificial procedure, 

 but it has important analogies with the process of natural 

 infection. Further on (p. 25) I shall mention Fenner's evidence 

 that natural infection of the rabbit with myxomatosis is often, 

 or usually, with a single infective particle. 



Even with mass transfer of culture, circumstances may 

 result in an elimination of variants in a fashion almost 

 analogous to what takes place in single colony isolation. In 

 Atwood, Schneider and Ryan (1951^, b) the behaviour of 

 a histidine-requiring strain of E. coli h~ and the wild-type 

 revertant A+ has been studied by transferring 0-5 ml. of old 

 culture to 50 ml. of new complete culture medium at daily 

 intervals. In this medium there is no selective advantage of 

 either type and appropriate experiments indicate a mutation 

 rate of h~ to h+ of 2-7 x io~^ and of A+ to h~ of i*2 x io~®. 

 In time, therefore, one would expect a corresponding equi- 

 librium ratio to be reached at 44A- to ih+. In fact, however, 

 the equilibrium situation is around 1*3^+ to lo^h-. Analysis 

 indicates that the discrepancy in favour of A~ is due to the 

 process of periodic selection. The only thing that matters for 

 survival in this or any other situation is an ability to produce 



II 



