222 THE ANTIGEN-ANTIBODY REACTIONS 



valent haptens do not precipitate with homologous antibody, though they can 

 be shown to combine with it. It is argued (Marrack 1938, PauUng 1940) that 

 if antibody is divalent, divalent hajDtens should aggregate with it in long chains ; 

 with tri- or multi-valent antibody, divalent haptens could form a lattice and 

 precipitate. 



Precipitation of divalent haptens was observed by Landsteiner and van der Scheer 

 (19521)) and by Pauling, Campbell and Pressman (1941) ; and Pauling, Pressman and Ikeda 

 (1942) showed that the hapten-antibody ratio was constant throughout the titration 

 range of a divalent hapten with homologous antiserum. The observed molar ratio was 

 0-75, as against the ratio of unity to be expected if both the antigen and the antibody 

 were bivalent. Hooker and Boyd (19416) and Boyd (1942) obtamed no evidence of pre- 

 cipitation or of the formation of long chains of alternating antigen and antibody. Land- 

 steiner and van der Scheer's haptens may have polymerized in solution and so formed 

 multivalent hapten particles ; though there is no positive evidence that the same criticism 

 is appUcable to the results of Pauling and his co-workers. Boyd (1942), from a study 

 of specifically precipitable trivalent arsonic haptens, concluded that their jjrecipitability 

 depended, not on their power to form a lattice, but on the closeness of their combinuig 

 groups to one another. When the arsonic combining groups were attached to a compound 

 so that they were separated by a relatively large molecule, the hapten was not precipitable 

 by antibody. Boyd suggests that after the arsonic groups of this latter type of hapten 

 have united with antibody, the number of hydrophile groups remaining free is sufficient 

 to keep the compound in solution. When these groups were rendered less hydrophilic 

 by acetylation, the haptens became precipitable. Precipitation on this basis is due 

 to the lowering of solubility by mutual neutralization of polar groups in antibody and 

 napten, and by crowding of antibody molecules together so tliat other polar groups are 

 occluded. This " occlusion " hypothesis is an extension to haptens of the " two-stage " 

 theory as applied to the reaction of full antigens in antibody excess. The serological 

 reactions of haptens, however, are not strictly parallel to those of full antigens, and this 

 lack of parallelism may invaUdate any conclusions that we may draw about the antigen- 

 antibody reactions. For example, Woolf (1941), working with antigens synthesized by 

 conjugating protein to hapten, observed that the antigen-antibody ratios were dependent, 

 as in other systems, on the amounts, and not on the concentrations of the reactants. 

 Inhibition of the reaction by hapten, on the other hand, depended on its concentration, 

 a fact which suggested that hapten-antibody compounds were far more dissociable than 

 antigen-antibody compounds. Moreover, Hershey (1942) points out that experiments 

 with haptens are unlikely to indicate anything about the valency of antibody, since 

 the discrepancy in size between the reactive groups of the haptens and a reactive patch 

 (i.e. unit " valency ") on the antibody is such that a modified lattice will form in the 

 early stage of precipitation, whether antibody is monovalent or not. In a later paper 

 (1944) he shows on theoretical grounds that the combination of antibody with divalent 

 hapten would not result in specific precipitation, since it is highly unlikely that sufficiently 

 long chains of alternating antigen and antibody molecules could form ; precipitation in 

 such circumstances would imply either a non-specific effect, or irreversible antigen -antibody 

 linkages ; or, as already noted, either an increase in the valency of hapten by its poly- 

 merization, or an effective multivalence of unit reactive patches on the antibody molecule. 



It is impossible at present to decide between the lattice and the two-stage 

 hypotheses, at any rate as mutually exclusive hypotheses. The proponents of 

 both are united in assuming that the initial combination is determined by specific 

 chemical factors, and that antigens are multivalent. It should be noted that the 

 demonstration of the multivalence of antibody will not alone prove that a specific 

 lattice is formed ; it will be necessary also to show that the antigen-antibody 

 linkages are specific, and that the bonds between the molecules are reversible 



