EHRLICH'S SIDE-CHAIN THEORY 199 



As an instance we may cite the quantitative results obtained in the neutralization 

 of toxin by antitoxin (see pp. 238-240). These could not be reconciled with the side- 

 chain theory in its simple, unmodified form. In order to retain the conception of 

 firm union between antigen and antibody in constant proportions, Ehrlich found it 

 necessary to postulate the existence of a large number of different toxin components 

 with varying degrees of affinity for antitoxin. Similarly, some of the results 

 obtained in his studies on hsemolysins made it necessary to assume the intervention 

 of receptors of considerable structural complexity. The consequent coining of a 

 host of new terms, designating components the actual existence of which was 

 extremely doubtful, served to confuse the problem rather than to clarify it. In 

 spite of these failings, Ehrlich's theory has the outstanding merit that it has kept 

 firm hold on chemical specificity as the essential feature of the antigen-antibody 

 reactions. If we compare his picture of the activities of the living cell with 

 the hypothesis of cellular metabolism referred to on p. 60 we shall note obvious 

 points of similarity ; though the studies of recent years have largely modified the 

 cruder and more hazy concepts of cellular processes on which the side-chain theory 

 had of necessity to be built. The strains to which this theory was subjected arose, 

 in the main, from the effort to retain the conception of firm chemical union in 

 constant proportions. We shall see later the direction in which an escape from 

 this dilemma may be found. Meanwhile we may turn to other theories that have 

 had an important influence on the development of our present views. 



An alternative theory, propounded by Bordet (1899, 1903), denies the appli- 

 cability of the laws of ordinary chemical union to reactions between antigens 

 and antibodies, and regards them as belonging to the class of colloidal reactions, 

 in which the determining conditions are physical rather than chemical. On this 

 view the laws which describe the phenomena associated with adsorption at sur- 

 faces and interfaces should be found to hold true, when antigens interact with 

 antibodies. The supporters of this hypothesis have been energetic opponents of 

 the view put forward by EhrUch and his school, and the output of immunological 

 literature during the first twenty years of the century was largely concerned with 

 this controversy. Bordet's theory makes no attempt to present a complete picture 

 of the mechanism underlying the serum reactions. He himself insisted that 

 his conclusions are so little removed from the facts observed, that they scarcely 

 merit the name of hypotheses. He quite frankly leaves many observed phenomena 

 unexplained, contenting himself with the claim that all the available data are 

 consistent with the limited views he has expressed, while pointing out inconsistencies 

 between these data and the more detailed conceptions of the supporters of the 

 side-chain theory. The essential simplification introduced by Bordet's hypothesis 

 was that, in bringing antigen-antibody reactions into the category of colloidal 

 phenomena, it abandoned the necessity for assuming combination in constant pro- 

 portions. One colloid may combine with another in proportions that vary over a 

 wide range. This advantage was gained at the initial expense of leaving specificity 

 out of account ; but more recent studies of the factors that determine adsorption 

 at surfaces and interfaces (see pp. 263-265) have gone far towards filling this 

 gap. 



Arrhenius (1904, 1915), alone and in co-operation with Madsen (1902, 1904), 

 has attempted to retain the advantages of the chemical basis of Ehrlich's theory, 

 while avoiding the necessity of making a fresh assumption to meet each new fact 

 observed. He has suggested that the reaction of an antigen with its antibody is 



