GENETIC INTERACTIONS BETWEEN ANIMAL VIRUSES 277 



single virus particle from whose progeny two or more genetically distinct and 

 true breeding clones can be obtained. To satisfy this criterion experimentally, 

 it is necessary to show that, from the population presumed to contain 

 heterozygotes, an undue proportion of hosts receiving an average of less than 

 one 50 % infective dose, allows the development of virus of two or more types. 

 In practice, this has always meant the use of serological methods to detect the 

 presence of two different parental serotypes in infective material from limiting 

 infective dilutions. 



When a double infection is produced with two different serological types of 

 influenza virus in the allantoic cavity, the fluid obtained will often show the 

 phenomenon of "double neutralization" of hemagglutinin. This indicates that 

 the surface of individual particles must be made up of a mosaic of the two 

 parental antigens. Hirst and Gotlieb (1953) use the term phenotypic mixtures 

 for virus particles of this character. It is reasonable to believe that a large 

 proportion of the viable particles in such populations are heterozygous, but 

 since the two characters of double neutralization and heterozygosity are 

 demonstrable respectively in large populations only and in individual 

 infective units only, it is impossible to provide a rigid proof of this. 



It seems highly probable that a fourth type of genetic interaction exists by 

 which unstable forms emerge which cannot be maintained as clones of 

 constant properties. This applies particularly to the virulence characters. 

 When a new virulence is transferred by recombination to another serological 

 subtype, it tends to be of lower level and more unstable than in the parent. 

 We have therefore spoken of "redistribution of virulence" as something not 

 quite the same as true recombination (Burnet and Lind, 1954b). 



III. The Technical Requirements in Virus Genetics 



A. Pure Clone Isolations 



As in comparable work with bacteria, all genetic work with viruses should 

 ideally be carried out with pure clones derived from a single parental particle. 

 This is rendered almost impossible in all practical situations because of the 

 fact that even with the best techniques for maintaining viability most of the 

 viruses which have been studied show a significantly greater number of 

 morphological particles than of viable units. With influenza viruses, standard 

 allantoic fluids contain about 10 morphological to 1 infective unit (Donald and 

 Isaacs, 1954). Horsfall (1954) contends that with adequate technical care the 

 ratio can be reduced to unity. Even with the most elaborate care, polio- 

 viruses show a ratio of the order of 30-100 : 1 and with virus produced in 

 standard fashion 1000-2000 morphological units may be present for each 

 infective (plaque-forming) unit (Schwerdt and Fogh, 1957). 



