150 ANTIGENS AND THE TECHNIC OF SERUM REACTIONS 
by neutral salts, and the precipitation of finely suspended clay by the 
addition of neutral salts; the inference has been drawn that the phe- 
nomenon of agglutination is one of physico-chemistry.^ The highly 
significant work of Northrop and DeKruif^ has shown that the two 
principal factors involved in the agglutination of bacteria are: first, the 
repellent force, due to the electric charge on the microbes themselves, 
which tends to keep them apart, and second, the cohesive force, which 
is impressed upon the organisms when they are bathed with specific 
sensitizing immune serum, and which tends to hold them together once 
they have collided. It follows that substances which reduce the 
surface charge of bacteria below a certain level will cause clumping, 
unless the cohesive force is also reduced. 
Specificity of Agglutination Reactions: Group Agglutinins.— The 
composition of the agglutinogen— that constituent of the bacterium 
which stimulates agglutinin formation— is unknown, but it appears 
to be complex and probably not a single chemical compound. Closely 
related bacteria may possess in common a small amount of agglu- 
tinogen— a least common multiple, as it were— which stimulates the 
production of "group agglutinin" that reacts with related bacteria 
more or less in proportion to their content of the common antigen or 
agglutinogen. The specific agglutinin produced by the entire agglu- 
tinogen content of an organism is more potent and fails to react with 
related bacteria. Thus, the serum of an animal immunized against 
B. typhosus may agglutinate that organism in a dilution of 1 to 3000; 
B. paratyphosus will be agglutinated in a dilution of perhaps 1 to 
300 by the same serum, and B. coli would agglutinate possibly only in a 
dilution of 1 to 50. The group agglutinin in this example would be effec- 
tive for B. paratyphosus in a dilution of 1 to 300, but in greater dilu- 
tions it would be ineft'ective. For B. coli in the instance cited, the 
group agglutinin is ineffective in dilutions above 1 to 50. 
The common or group agglutinin for B. paratyphosus in this typhoid 
serum could be quantitatively removed by leaving it in contact with 
a large number of paratyphoid bacilli for a few hours, then centrifu- 
galizing to remove the organisms. The residual serum would contain 
only agglutinin specific for B. typhosus. If B. typhosus were added 
to the serum, all the agglutinin— both "group" and specific— would 
be removed. 
As a general rule, group agglutinins constitute a minor fraction of 
the total agglutinin and in practice the degree of dilution of the serum 
used in specific cases is ample to exclude error. It occasionally 
happens that sera of low dilution, especially those rich in agglutinoids, 
fail to clump the specific organism; as the serum is diluted more and 
more the phenomena of clumping become more and more marked; 
finally a degree of dilution is reached beyond which the serum again 
becomes ineffective. The initial negative agglutination in concen- 
1 See Phillip: Phys. Chem. 2d ed., 1913, pp, 216, 240. 
2 Northrup and De Kruif: Jour. Gen. Physiol., 1921-1922, 4, 639, 664. 
