VARIATION IN VIRULENCE 231 



passage, either in the amnion at low dilution or in the allantois, the virus 

 acquired the capacity to agglutinate fowl cells to as high a titer as human 

 cells. This change, from the ("original") to the D ("derivative") phase, is 

 associated with a change in affinity for various hemagglutinin inhibitors 

 (Stone, 1951) and in position in various receptor gradients (Burnet et al., 

 1946a, 1949); it is also associated with a marked reduction in virulence for man 

 (Burnet and Foley, 1940) and emergence of capacity to multiply in the allantois 

 (Burnet and Bull, 1943). 



However, it has been possible to maintain virus in the phase through 

 repeated passages in the amnion if each passage is initiated at "limit dilution" 

 (Burnet and Bull, 1943). This shows that a mere history of prolonged multi- 

 plication in the amnion is not in itself a certain inducer of the change to D 

 phase: for this to happen there has to be full opportunity for minority 

 populations to supplant the original type. It is as if D phase virus arises 

 relatively infrequently during multiplication in the amnion; although 

 endowed with considerable survival advantages over phase virus, D phase 

 virus must seldom represent the majority type after a single passage in the 

 amnion. These considerations led Burnet to propose that the O-D change 

 represents a relatively rare mutation backed by intense selection pressure, 

 rather than a host-induced change. As we shall see later, an exactly analogous 

 argument has been proposed as evidence for the mutational nature of the 

 attenuation of poliovirus in tissue culture (Sabin et al., 1954). 



Unfortunately, the O-D change, on more detailed examination, proved to 

 be more complex than originally supposed. The dividing line between 

 agglutinative behavior of 0- and D phase is not completely clear-cut. Between 

 the two extremes (no agglutination of fowl cells, even in the cold, and agglutina- 

 tion of fowl and human cells to equal titer) there occur intermediates which 

 are not easily distinguished from mixtures (Burnet and Stone, 1945a). 

 Continued passage of phase virus at limit dilution in the amnion tends to 

 exaggerate the character of its agglutination, so that strains which showed 

 slight agglutination of fowl cells when originally isolated, show no agglutina- 

 tion of fowl cells after prolonged passage in the amnion; perhaps it was because 

 of this that the definition of phase agglutination became progressively more 

 stringent with the passage of time (Burnet and Bull, 1943; Burnet and Stone, 

 1945a; Burnet, 1950, 1951; Burnet et al., 1949). At least one strain, IAN, has 

 stabilized in the phase on passage in the amnion so that D mutants are 

 not readily obtained (Burnet et al., 1949); there is at least one instance of 

 conversion of D phase to phase by prolonged passage in the amnion (Magill 

 and Sugg, 1948). Further, D phase virus (or intermediate phase virus) does 

 not necessarily possess survival advantages over phase in the amnion, since 

 the only experiment testing the behavior of artificial mixtures in the amnion 

 gave inconclusive results (Burnet and Stone, 1945a). 



