274 THE ANTIGENIC STRUCTURE OF BACTERIA 



hsemolytic antibodies. When an agglutinating serum is exposed to an excess of 

 bacterial cells that contain some or all of the corresponding agglutinogens the 

 agglutinins that find their counterparts in the antigenic structure of the bacteria 

 are bound, or absorbed, and thus removed from the fluid in which the bacteria 

 are suspended. If the bacteria and the antibodies adsorbed to them are separated 

 by centrifugation, the supernatant fluid can be tested as regards its remaining 

 agglutinins. For instance, taking the illustrative example given above, if Serum 1 

 were absorbed with an excess of Bacterium 1 all the agglutinins A, B, C, D, and E 

 would be removed and the supernatant fluid would have no agglutinating action. 

 If it were absorbed with Bacterium 2, C, D and E would be removed, but A and 

 B would remain. The supernatant fluid would not agglutinate Bacterium 2 — 

 the absorbing strain — nor Bacterium 3 (since that bacterium, does not contain 

 agglutinogens a or 6) ; but it would still agglutinate Bacterium 1 in virtue of the 

 remaining agglutinins A and B. If it were absorbed with Bacterium 3, the single 

 common agglutinin E would be removed. The supernatant fluid would (as always) 

 fail to agglutinate the absorbing strain ; but it would still agglutinate Bacterium 1 

 in virtue of agglutinins A, B, C, and D, and Bacterium 2 in virtue of agglutinins 

 C and D. 



A note may be interpolated here with regard to two terms that are sometimes loosely 

 applied. The antiserum that is produced by the inoculation into a suitable animal of a 

 particular bacterium is frequently referred to as a hotnologous serum. A serum that 

 agglutinates the same bacterium but has been produced by the inoculation of some other 

 bacterium, dififering in one or more of its' antigenic components, is termed a heterologous 

 serum. Used in this sense the terms are useful and logical, and may be appUed either 

 to an antiserum in relation to a bacterium, or to two antisera or two bacteria in relation 

 to one another, implying in either case complete correspondence between active combining 

 groups. Thus, a serum that contains the active antibody groupings A, B, C, D is homo- 

 logous with a bacterium containing the active antigenic groupings a, b, c, d ; two bacteria 

 each containing the active antigenic groupings d, e, f, g are homologous with one another, 

 and so on. But an antiserum contahiing the active groupings A, B, C, D is not homo- 

 logous, either with a bacterium containing the active groupings a, b, c, e, or a, b, c, d, e, 

 or a, b, c. The terms cannot, logically, be apphed to smgle antigens or single antibodies. 

 They describe the relation between a group of different antibodies in a serum and a group 

 of antigens attached to a bacterial cell, or between the group of antigens attached to 

 one bacterial cell and the group attached to another. We are also, it should be noted, 

 using the terms quahtatively, not quantitatively. 



Suppose that we have an unknown bacterium x, which we suspect to be antigenically 

 homologous with a known bacterium y, and suppose that we have available an anti-i/ serum. 

 We absorb the anti-y serum with x and then test it against y. If it now fails to agglutinate 

 y we may assume that all the active antigenic groupmgs in y are also present in x. But 

 we have not excluded the possibUity that x has additional active antigenic groupings, not 

 present in y. So we prepare an anti-a; serum, and then absorb it with y. If, after absorp- 

 tion, it fails to agglutinate x, we assume that all the active antigenic groupings present 

 in X are also present in y. That is, x and y are antigenically homologous. This double 

 cross-absorption method is often spoken of as the " mirror test." 



The results of agglutination and adsorption tests with related strains of bacteria and 

 their respective antisera, when interpreted in the light of these principles, have in many 

 cases provided detailed pictiu-es of the antigenic make-up of certain bacteria. These 

 pictiu-es are for the most part quahtative, though it is usually possible to guess roughly 

 which of the antigens predominate. The vahdity of these quahtative interpretations 

 depends on the assumption that the separate antigenic components react only with 



