POPULATION GENETICS 



they are progressively less likely to take on the filamentous 

 form (Burnet and Lind, 1957). 



Our experience can be summarized as follows. Limit- 

 dilution passage will maintain a virus in its original form, 

 despite passage under conditions where potential mutants 

 could outgrow it, only if two conditions are fulfilled, {a) The 

 dominant form of virus in the initial population is homozygous 

 at least in regard to all factors which would influence survival 

 in the new environment, {b) Mutation does not occur with 

 sufficient frequency to overgrow the original clone before 

 harvesting, or alternatively, when this does occur its mani- 

 festation is clear enough to allow the elimination from the 

 passage series of the preparation in question. 



{d) The process of viral infection 



Both in the laboratory and in the field the process by which 

 a viral infection is initiated is liable to be a complex one, very 

 subject to interruption by essentially accidental circum- 

 stances. The standard approach to this question is to examine 

 the end of the titration curve to see whether the proportion 

 of infections produced by dilutions fairly closely spaced about 

 the ID50 level correspond to a Poissonian distribution. The 

 general finding is that, in fact, results do correspond fairly 

 closely, even if we know by experiment that each infective 

 dose in some instance corresponds to many hundreds or 

 thousands of infective doses for some more susceptible host 

 (Parker, 1938). In these circumstances, we have always the 

 alternative to consider, (a) whether the population of virus 

 particles contains one in each thousand which, because of 

 some intrinsic physiological or genetic difference from the 

 other 999, can alone initiate infection, or {b) whether most or 

 all of the 1 000 particles can initiate infection but the chance 

 that any one will do so is approximately i to 1000. Perhaps 

 the clearest indication that the second alternative has at least 

 something to do with the situation is contained in Cairns' 

 (1957) work on the asynchrony of influenza- virus infection. 



21 



