SEROLOGICAL VARIATION 259 



the 29 strains studied, 18 different antigens were detectable, and as many as 

 15 could be demonstrated in a given strain. Takatsy and Furesz (1954) have 

 raised some technical objections to the procedures from which these conclu- 

 sions were drawn. They accept, however, the principle of common antigens. 

 The strains tended to group chronologically according to major antigens. 

 The older strains exhibited limited crossing with strains of later years, but 

 the latter frequently exhibited numerous common components with strains of 

 the earlier years. It is clear that an adequate view of antigenic composition 

 and relationships is obtained only by testing each virus against numerous sera 

 reflecting the composite antigenic make-up of the individual strains. The 

 variation in strains from different epidemic years or periods can thus be seen 

 in relation to the antigenic spectrum as a whole. When strains from different 

 periods were compared by the same procedure, it was noted that, despite 

 individual variations, strains from 1933 to 1943 exhibited distinct similarities 

 and that the deviations among the group were chronologically random rather 

 than showing a directed mutational shift in the accession of new antigens with 

 succeeding years. The same situation obtained with a representative series of 

 strains isolated in the years 1946 to 1955. Strains of swine influenza virus 

 over a 23-year period were, with two exceptions, essentially identical (Jensen 

 and Peterson, 1957). 



d. Interpretation. The substance of these analyses is to emphasize the 

 presence of numerous antigenic constituents common to a large proportion of 

 strains of influenza virus type A; the dominant antigen of one may be so 

 much reduced in another as to be detectable only by procedures which 

 disclose the full antigenic complement. While one component becomes 

 enhanced as a characteristic strain antigen, others may recede to represent 

 only secondary or group antigens. Variation in strains is viewed then as 

 primarily a process of quantitative or spatial rearrangement, rather than 

 complete loss of existing antigens and the development of completely new 

 antigenic structures. Rearrangement of 18 antigens of itself permits an almost 

 inexhaustible supply of variants. The bulk of evidence is in support of this 

 thesis, but some investigators prefer the loss and gain motif (Andrewes, 

 1954, 1957). 



The appearance of the 1957 Asian variant brings another abrupt transition 

 which, on the basis of present information, suggests a sharp mutational shift, 

 bearing little similarity to recent A-prime strains or to isolated strains from 

 earlier years. Studies of its antigenic character are not sufficiently advanced to 

 portray its make-up, but observations with human serum demonstrate that it 

 is related to strains of earlier distribution. 



Hilleman (1954) has pointed out that in these efforts toward analysis and 

 classification the procedures employed and the materials used have differed so 

 widely as to make interchange of the data difficult. Certain generalizations 



