258 T. FRANCIS, JR. 



(Mulder et al., 1956), although their results continue to show cross relation- 

 ships at low levels. Takatsy and Hamar (1955) insist that the 1953 strains are 

 direct mutational descendants of a 1952 strain. 



c. Analysis by .Antibody- Absorption. Friedewald (1944) first employed 

 antibody-absorption methods effectively in antigenic analysis of 6 influenza 

 strains. Concentrated virus from allantoic fluid was used to absorb convales- 

 cent ferret and convalescent or postvaccination human serum. The effect on 

 homologous and heterologous antibody was measured in cross tests by mouse 

 neutralization, HI, and CF methods. Each strain removed all antibody from 

 its homologous serum and varying amounts of cross-reacting antibody from 

 the different heterologous sera. Clear differences and relationships between 

 the test strains were demonstrable. It is of interest that the PR8 strain 

 removed antibody against all strains from convalescent or vaccinated human 

 serum at that time, while other strains exhibited more strain-specific effect. 

 The patient's own strain also removed all antibody from his convalescent 

 serum although it differed somewhat from PR8. Antibody as determined by 

 all three serological methods was equally affected. 



Hirst (1952) reported antigenic grouping by antibody absorption of 61 

 strains covering the 18 years from 1933-51. He employed hyperimmune 

 rabbit serum against swine 1931, WS 1933, PR8 1934, Melbourne 1935, 

 Talmey 1937, Alabama 1941, and FM X 1947 strains. Each serum was absorbed 

 with heterologous strains until only antibody to the homologous prototype 

 virus remained. Each strain was tested by HI against the 7 strain-specific 

 sera. This represented, therefore, largely a one-way grouping. Swine, WS, and 

 PR8 groups appeared distinct, while strains from 1935-37 were divided 

 among Melbourne, Talmey, and Alabama 41 groups. The latter also included 

 all strains from 1939 to 1944; the strains from 1946 to 1950 were in the FM X 

 group. The strains of 1950-51 were separate. Certain strains from different 

 periods were not classifiable. 



Antigenic analysis was extended by reciprocal serological reactions with 

 sera absorbed through periodate-treated virus fixed to formalin-treated cells 

 (Jensen and Francis, 1953). Theabsorptive capacity of each cell- virus complex 

 was standardized against homologous immune serum and equal absorptive 

 amounts were then used for comparative observations. Sera of ferrets 

 convalescent from infection with one of 18 strains were pooled in quantities 

 which provided a pool with equivalent HI titers against each strain. Absorp- 

 tion with a given strain removed antibody to itself, whether the antibody 

 derived from homologous serum or from cross-reacting antibody from sera 

 against other strains. It reduced the titers to heterologous strains presumably 

 proportionate to the amount of the same antigens shared by it and hetero- 

 logous strains represented in the antibody pool. Thus, a profile of the quantita- 

 tive and qualitative antigenic composition of a strain was obtained. Within 



