262 THE ANTIGEN -ANTIBODY REACTIONS 



(see Stull and Hampton 1941). Landsteiner (1942) found that polypeptide chains of 

 8-12 amino-acids from hydrolysed silk fibroin were haptens in that they inhibited the 

 reaction between fibroin and specific antibody, and concluded that the determinant units 

 of the protein were of this size. 



A number of mechanisms can be used by an animal to remove or neutralize 

 a substance that has been introduced into it ; such as excretion, if the substance 

 is small enough to get through the kidney ; hydrolysis, if suitable enzymes are 

 present ; absorption by phagocytes or by the cells of the reticulo-endothelial 

 system (see Chapter 50) ; or, in the case of substances like insulin, whatever 

 mechanisms normally remove indiffusible hormones, etc., after they have produced 

 their physiological effect. In these circumstances a substance will be antigenic 

 only if it is too big to be excreted, too stable to be hydrolysed and so unusual 

 that it is more readily handled by the mechanism resulting in antibody-formation 

 than by the more direct physiological modes of destruction and elimination. 



In summary, then, it appears that to be an antigen a substance must have 

 certain minimum chemical properties, but it must also be so constructed that 

 it cannot be handled by one of the readily available mechanisms of elimination, 

 and thus removed before it has time to exert an antigenic stimulus. 



The Effective Number of Determinants, and the Valency of Antigen. 



The valency of native antigens may be estimated from analysis of specific 

 precipitates in the region of antibody excess. This figure gives only the minimal 

 number of combining groups on the antigen molecule, since spatial considerations 

 may not permit their full saturation with antibody in the precipitate. The values 

 range from 5 for crystalline ovalbumin (molecular weight about 42,000) (Heidel- 

 berger and Kendall 1935c, Heidelberger 1938) to 231 for the hsemocyanin 

 (molecular weight 5 — 6 X 10^) of the crab Viviparus (Malkiel and Boyd 1937). 



The number of valencies is correlated with the surface area of the antigenic 

 protein molecules (Hooker and Boyd 1942), and it appears that the maximal size 

 of a determinant, supposing it to be protein in nature, is of the order of 30 amino- 

 acids. 



By coupling arsanilic acid in varying proportions to casein, and testing the 

 compounds against antisera prepared against another arsanil-azo-protein. Hooker 

 and Boyd (1932) calculated that an average of at least 13 introduced groups per 

 molecule of casein was necessary for precipitation. Haurowitz, Kraus and Marx 

 (1936) similarly estimated 10-20 groups for arsanil-azo-globulin. The assumption 

 is made that the protein particles are of full molecular weight. During the prepara- 

 tion of azo-proteins, however, the average particle size may decrease (Hooker, 

 personal communication) and the minimal number of reacting groups may in fact 

 be lower than the calculated 10-13. 



Unrecognized Antigens. 



Our definition of an antibody implies that it should react in a detectable manner 

 with the corresponding antigen, and we consequently infer antigenicity when a 

 substance reacts specifically in a detectable manner with the " antiserum " pre- 

 pared against it. It is nevertheless possible for a substance to be antigenic in the 

 formal sense of stimulating antibody production, and yet be non-reacting. As 

 we saw in the preceding section, 10-20 arsanilic acid determinants were neces- 

 sary for the precipitation of arsanil-azo-protein by its antibody. Yet the intro- 

 duction of an average of one arsanilic acid group per molecule of protein gives 



