IMMUNOLOGICAL METHODS IN THE STUDY OF VIRUSES 545 



Two theoretical interpretations have been published to account for the 

 empirical relationship 



alog>S=logFo-logFi+6 



where a is the constant determining the slope of the curve and b is equivalent 

 to the reduction that would be shown by undiluted serum. 



Tyrrell (1953) assumes that in the allantoic cavity a higher-order reaction 

 is needed to convert a just unneutralized particle to an inactive one and that 

 a is a measure of the number of molecules of antibody needed. Dulbecco et al. 

 (1956), on the other hand, believe that the process cannot be defined by the 

 initial concentration of the reagents. They believe that in the embryo where 

 virus is just neutralized, primary interaction of virus and antibody allows 

 infection of a jDroportion of cells. These liberate virus of which a high propor- 

 tion is neutralized by residual antibody, but a small proportion can initiate 

 a second cycle of cell infection. Whether or not infection results, as measured 

 conventionally by the appearance of hemagglutinin at 3-4 days, will depend 

 on whether successive cycles give progressively smaller yields or whether at 

 some point a virus breakthrough occurs. Dulbecco interprets a in the formula 

 as a measure of the average number of generations before "cut-off" in virus 

 multiplication occurs. 



Another aspect of virus neutrahzation exempUfied by influenza viruses 

 may be of great theoretical importance, although its imphcations have not 

 been widely discussed. Partial absorption of an immune serum by virus 

 reduces antibody as measured by some tests much more than by others. This 

 was shown first by Burnet et al. (1937) who compared titrations on the 

 chorioallantois and in the mouse lung. Using a strain of Melbourne patho- 

 genic, both for mouse lung and for chorioallantois, it was shown that the 

 same mixtures uith absorbed antiserimi would indicate a 99 % removal of 

 antibody by the chorioallantoic route, but a retention of one-third when tests 

 were made in mice. Walker and Horsfall (1950) rather similarly showed that 

 an absorbed antiserum could lose most of its activity as tested by mouse lung 

 neutralization but retain it in antihemagglutinin tests. 



6. Conclusion 



The over-aU impression that one receives from both the quantitative and 

 quahtative studies of animal virus neutrahzation is the heterogeneity of both 

 antibody and virus populations. Influenza studies may not be wholly relevant 

 to those with smaller viruses but it is reasonable to expect that when equally 

 extensive methods are available the same complications wDl emerge with 

 forms like pohovirus and WEE virus. If this is so, the only useful generaliza- 

 tions are hkely to continue to be expressed in quahtative terms. 



