THEORIES OF IMMUNITY 127 
tively thermostabile. The non-specific complement (found in fresh 
blood serum from any animal) is thermohibile and readily destroyed. 
Thus far it has been assumed that the cells of the body defend 
themselves against toxins, alien protein or alien cells by the forma- 
tion of specific antibodies or side-chains. Welch^ has made the 
important suggestion, which has experimental evidence in its favor, 
that bacteria may also produce side-chains which are specific for 
certain cells of the host. A struggle between host and microbe, 
therefore, would not be one-sided; a dual attempt at immunization 
is going on during a bacterial invasion, in which the microbe attempts 
to protect itself against the specific weapons of the host as the host 
attempts to protect itself against the weapons of the in\ading micro- 
organism. Thus, bacteria grown in media containing agglutinating 
sera gradually lose their agglutinability, but this acquired loss of 
agglutinating power is not exhibited by descendants of the inagglu- 
tinable strain grown for some time in media not containing agglutinins. 
The side-chain theory, originally formulated to explain antitoxin 
immunity, but enlarged in its scope to include the phenomena of 
agglutination, precipitation and cytolysis, has been subjected to 
much adverse criticism. It was assumed by Ehrlich at first that toxin 
and antitoxin, for example, united in simple proportions as a strong 
acid and a strong base unite; the chemical analogy of toxin-antitoxin 
union to form an inert mixture comparable to a salt was further accent- 
uated by the effect of moderate degrees of heat in hastening the reaction 
between the two. A very thorough investigation of the quantitative 
neutralization of toxin by antitoxin revealed the error of this supposi- 
tion and P^hrlich was subsequently led to assume a very complex struc- 
ture for the toxin molecule, in which there existed several fractions 
possessing individually, diflferent affinity for antitoxin. 
Madsen and Arrhenius^ studied the toxin-antitoxin union from 
the standpoint of physical chemistry and found that the slightly dis- 
sociated reactive substances united in conformity with the law of 
mass action of Guldberg and Waage. Their conclusion was that 
toxin and antitoxin react like a weak acid and weak base, and that 
it is a reversible reaction, so that a mixture of toxin and antitoxin 
always contains free toxin, free antitoxin and toxin-antitoxin, the 
relative amounts being calculable according to the law of mass action. 
The observations of Theobald Smith'^ and of many other observers 
that seemingly neutral mixtures of toxin and antitoxin would induce 
active immunity in experimental animals are in harmony with this 
view. Biltz, Much and Siebert* have advanced an hypothesis, based 
upon the assumption that toxin and antitoxin are colloids, which in 
essence assumes that the toxin-antitoxin reaction is a phenomenon of 
' Huxley Lecture, British Med. Jour., 1902, ii, 1105. 
2 See Arrhenius: Immunochemie, Leipzig, 1907, for full details. 
' Active Immunity Produced by So-called Balanced or Neutral Mixtures of Diphtheria 
Toxin and Antitoxin, Jour. Exper. Med., 1909, 2, 241. 
* Behring's Beitrage zur exp. Therapie, Heft. 10. 
