232 F. FENNER AND J. CAIRNS 



In fact, the 0-D change does not really lend itself to a detailed examination 

 of the processes of mutation and selection, since there is no clear-cut procedure 

 for assaying separately either or D phase virus. Originally it was supposed 

 that only D phase virus multiplies in the allantois, which therefore could be 

 used as a specific host for assay of D phase virus. However, phase virus has 

 been shown to multiply in the allantois to a limited extent (Burnet and Stone, 

 1945b), so that there is no precise procedure for distinguishing mixtures of 

 and D phase virus from intermediates. In view of these difficulties it is 

 clear that no accurate determination of the relative survival advantages of 

 the various types can be made, and hence no estimate of mutation frequency 

 is possible. 



Lastly, the idea that agglutinative behavior bears any simple relation to 

 the host range or past history of any particular strain is itself difficult to 

 sustain. Influenza B strains are not in the phase on first isolation (Burnet et 

 al., 1946b; Hirst, 1947a); indeed, on passage, the agglutinative behavior of B 

 strains tends to proceed in the reverse direction (Ledinko and Perry, 1955). 

 Most recent influenza A strains have on isolation shown qualities closer to 

 D than to phase. And we shall see, in the section on adaptation of influenza 

 virus to mouse lung, that changes in hemagglutinating behavior, although 

 often associated with the process of adaptation, are not an obligatory step in 

 the change in host range but a secondary phenomenon which may bear little 

 relation to the alteration in host range. 



2. Variation of Poliovirus in Tissue Culture 



The study of the d mutation by Dulbecco and Vogt (1958) has already been 

 discussed in the section of mechanisms of variation. It remains in this 

 section to deal with those cases where there is some relation between 

 neuropathogenicity and behavior in tissue culture. 



In general, prolonged passage in tissue culture results in a reduction of 

 neuropathogenicity. This, although not invariably true, has been demon- 

 strated repeatedly [Enders et al. (1952) and countless others since] during the 

 search for attenuated strains suitable as vaccines. There is some evidence for 

 the belief that this attenuation occurs by way of selection of spontaneous 

 mutants which have survival advantage in tissue culture and, incidentally, a 

 lowered neuropathogenicity. Thus, passage of Mahoney, Y-SK, and Leon in 

 cynomolgus kidney tissue cultures results in conversion to attenuated 

 variants only if large inocula are used for each passage; if passage is conducted 

 at "limit dilution" the strain keeps its neuropathogenicity, presumably 

 because one passage will seldom provide adequate opportunity for the 

 attenuated mutants to become the majority type (Sabin et al., 1954). This 

 result is exactly analogous to the findings with the O-D change (Burnet and 

 Bull, 1943). Alteration of neuropathogenicity to mice of a line of type 3 



