276 THE ANTIGENIC STRUCTURE OF BACTERIA 



comparison of bacterial strains in terms of the proportion of total antibody they 

 will absorb from a specific antiserum. Duncan (1932a, 6) and Miles (1933, 1939) 

 attacked the problem by means of the optimal proportions technique. The 

 equivalence point between a bacterial suspension and an antiserum may be the 

 resultant of the reaction of a number of different antigens and their corresponding 

 antibodies which may to some extent be distinguished by appropriate variations 

 of the experimental conditions. 



This leads us to the problem of differentiatmg two types of bacteria, both of which 

 contain the same antigenic components, but in different proportions. Using the qualitative 

 differential absorption technique, with maximal absorbing doses of bacteria, the two types 

 would appear identical, since each would absorb all the agglutinins from both the homo- 

 logous and heterologous serum, using those terms in this case to include quantitative as 

 weU as qualitative relations. By careful adjustment of the absorbing dose it is, however, 

 sometimes possible to demonstrate a quantitative diiference of this kind. An example, 

 dealing with the differentiation of Br. melitensis from Br. abortus, will be found on p. 824. 



The Study of Bacterial Variation as a Guide to Antigenic Structure. 



Bacterial variation, as a general phenomenon, is dealt with in the succeeding 

 chapter ; but the detailed study of the antigenic differences displayed by bacterial 

 variants has played so large a part in the formation of our present views in regard 

 to antigenic structure that we must anticipate a little and discuss in the present 

 section certain examples that will help to illustrate the kind of antigenic changes 

 that occur. 



Smith and Reagh (1903) isolated a non-motile variant of the hog cholera bacillus, 

 and compared its agglutination reactions with those of the normal, motile, flagellated 

 type of this organism. They found that a serum prepared against the motile^ strain 

 agglutinated both the motile and the non-motile strains ; but the titre was much 

 higher for the motile than for the non-motile, and the clumps formed by the motile 

 strain were fluffy and formed rapidly, while the clumps formed by the non-motile 

 strain were tight, small and granular, and formed far more slowly, A serum 

 prepared against the non-motile strain agglutinated the motile and non-motile 

 strains to the same degree, giving in each case the slow, granular type of agglutina- 

 tion. The serum prepared against the motile strain, when absorbed with the non- 

 motile strain, lost its power of agglutinating the non-motile bacilli, but retained 

 its power of agglutinating the motile strain. The serum prepared against the non- 

 motile strain lost its agglutinating power for both, strains when absorbed with the 

 motile strain. Smith and Reagh concluded that the normal, motile hog cholera 

 bacillus has two kinds of antigens, one contained in the flagella, the other in the 

 cell body. The non-motile type has lost the flagellar antigens and retains the body 

 antigens only. Beyer and Reagh (1904) extended these observations and showed 

 that the flagellar antigens were so altered by heating to 70° C. for 15 minutes that 

 the heated bacilli no longer gave the flagellar type of agglutination, but reacted like 

 the non-motile strain. The somatic, or body, antigens were not affected by this 

 treatment and still gave the characteristic slow, granular agglutination (see also 

 Orcutt 1924, Craigie 1931). 



These observations of Smith and Reagh and of Beyer and Reagh demonstrated 

 all the essential points of difference between the flagellar and somatic types of 

 agglutination, but they passed almost unnoticed until similar findings were recorded 

 by Weil and Felix (1917) in connection with their work on the diagnostic significance 

 of the agglutination of a particular strain of proteus bacillus by the blood of a patient 



