264 THE ANTIGEN-ANTIBODY REACTIONS 



that the structural differences between them do not exceed certain Hmits. Thus, 



and thiophene [ j form mixed crystals, as do 



for example, benzene 



azo-benzene 



-N=N- 



S 

 and stilbene 



— CH=CH 



The formation of mixed crystals might, perhaps, be regarded as analogous to 

 the formation of a precipitable antigen-antibody compound, particularly if we adopt 

 Marrack's lattice hypothesis. It is of interest to note that the phenomenon of 

 crystallization also presents an analogy to the inhibition of precipitation by haptens. 

 Many crystals will specifically adsorb on to their surfaces the molecules of another 

 substance, with which they will not form mixed crystals. Such adsorption inhibits 

 the growth of the crystal by the addition of further molecules of the substance of 

 which the crystal is composed. The conditions, in regard to molecular size and 

 position of active groupings, that determine the specificity of adsorption on the 

 surface of a crystal, are less strict than those that determine availability for building 

 into a complete crystal lattice. This would be expected, since only one aspect of 

 the adsorbed molecule need conform to the distribution of active groupings on the 

 surface of the molecule. Relatively small differences may, however, be sufficient 

 to determine the occurrence or non-occurrence of adsorption. Thus (France 1930) 

 the dye 



NH2 OH OH NH2 



NaO,S 



SO,Na 



NaO,S^ A jSOgNa 



is adsorbed on the cube faces of potash alum, while the dye 



NH, OH OH NH2 



NaO,Sr V ^— N=N-/^~N— /~V-N=N— /V^SO,Na 



NaOgS SOgNa 



is not. 



These analogies cannot be pressed far, for many of the substances which are 

 known to act as antigenic determinants could not find, in any of the geometrically 

 possible rearrangements of the polypeptide chains of an antibody globulin, con- 

 figurations that are specific in the sense that a surface of a crystal fits molecules 

 of its own kind. Moreover, Pauling, Campbell and Pressman (1943) point out 

 that the attractions between permanent electric charges and dipole moments 

 on the molecule are small in water, and have probably little effect in antigen- 

 antibody reactions; the attractions are more likely to be due to van der Waals' 

 forces, and to hydrogen-bonding. Van der Waals' forces are due to momentarily 

 induced charges between molecules having no permanent dipole moments. Con- 

 sidered singly, these forces are small, and would be strong enough to account 

 for attraction of antigen and antibody only if large areas of each molecule were 

 approximated sufficiently to give a summation effect of hundreds of van der Waals' 

 attractions. Hydrogen-bonding occurs when the hydrogen attached to a strongly 

 electronegative atom is attracted to the unshared electron pair of another electro- 

 negative atom (co-ordinate covalency). Since all these forces are non-specific, 



