268 T. FRANCIS, JR. 



increased by passage in certain tissues (Jensen et al., 1957), while passage in 

 mice may decrease susceptibility to the inhibitors. Another heat-labile effect 

 of serum which neutralizes virus infectivity was observed by Ginsberg and 

 Horsfall (1949) in mice. This has been studied in detail by Chu (1951) and 

 Tyrrell (1954). The active agency appears to belong to the class of natural 

 antibodies and is found in the a-globulins. Here again strains and lines may 

 vary considerably in susceptibility. Adaptation of strains to mice is associated 

 with development of resistance to this inhibitor, but the Asian 1957 variant 

 was initially refractory. 



Early observations with neutralization or active immunity tests showed 

 that strains varied in their reactivity with homologous and heterologous sera 

 (Magill and Francis, 1938; Francis and Magill, 1938). Hirst used the term 

 "avidity" for differences in antibody combining power among strains and 

 pointed to its influence upon serological definition of strains by hemagglutina- 

 tion technique. After systematic study, van der Veen and Mulder (1950) 

 classified strains in P, Q, and R phases according to their behavior in 

 hemagglutination-inhibition tests with homologous and heterologous ferret 

 sera. P phase virus is inhibited to high titer by its homologous serum only. 

 Q phase virus is poorly neutralized even by the homologous serum, while the 

 R phase is inhibited to high titer by homologous sera and by sera against 

 heterologous but related strains. They may be altered by passage in mice from 

 Q to P phase. Isaacs et al. (1952) reported change from P to Q phase when 

 immune serum was introduced into egg passage, but have also reported that 

 newly isolated strains may breed true in egg passage. The observations of 

 Jensen et al. (1957) have pointed out, however, that the nature of the host 

 tissue may be a distinct influence and phase alterations may in reality 

 represent quantitative variations. Isaacs (1953), however, proposed that the 

 P-Q difference might reflect changes in steric relationship of antigens, while 

 the observations of Fiset and Depoux (1954) with cross antibody absorption 

 led them to suggest that the differences are related to the presence of masked 

 antigens unavailable for antibody union and that these may be unmasked 

 selectively when passed through an immune host. This same unmasking may 

 conceivably result also from passage in an adequate animal host. The concept 

 is in keeping with the idea of antigenic ebb and flow within the virus particles. 

 The ready reversibility of the changes under a variety of conditions invites 

 speculation that they relate to quantity of surface lipids which binds the 

 specific antigens into a functional unit; Mulder et al. (1956) have speculated 

 that they may reflect some intervening biophysical property separate from 

 antigenic structure. Since these effects are primarily seen with ferret antisera, 

 not with chicken and mouse sera, the influence of such factors on serological 

 studies of antigenic classification may result in attention concentrated on 

 what are antigenic artifacts in mouse or egg lines rather than on true views of 



