236 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1950 



forces operating between atoms and small molecules fall off very 

 sharply with increasing distance, an effective bond would be formed 

 only if the two molecules were in contact with one another, that is, if 

 the surfaces of the atoms of antigen and antibody were no more than 

 a very few angstroms apart. The specificity of the bond formed in 

 this way would result from the detailed complementariness not only 

 in general surface configuration but also in the position of the groups 

 capable of forming hydrogen bonds and in the positions of the posi- 

 tive and negative electrical charges. It can readily be seen that this 

 mechanism does provide the possibility of very great specificity. Thus 

 a combining region with area of perhaps 200 square angstroms, rep- 

 resenting a surface of about 50 atoms, could be prevented from ap- 

 proaching to contact with the complementary region on the antibody 

 simply by replacing a methyl group, say, on the antigen surface by 

 a phenyl group, which would extend about 3 A. above the former 

 surface, and would hence hold the antibody 3 A. farther away from 

 the antigen, thus reducing the forces of attraction to such an extent 

 as no longer to permit them to result in a significant bond. 



The approximation of the antibody to the haptenic group of the 

 immunizing antigen must be very close. A striking bit of evidence, 

 from among the gi'eat amount that exists, is that of the cross reactivity 

 of two closely related haptenic groups, the m-aminobenzoic acid group 

 and the 4-chloro-3-aminobenzoic acid group, which differs from the 

 first only in having a chlorine atom in place of the hydrogen atom. 

 Landsteiner and his collaborators found that anti-4-chloro-3-amino- 

 benzoic acid serum precipitates readily both with the hapten-homol- 

 ogous azoprotein and with an azoprotein containing the ?n/-azobenzoic 

 acid group. On the other hand, the anti-m-azobenzoic acid serum 

 precipitates readily an azoprotein containing the m-azobenzoic acid 

 group, but does not form a precipitate with an azoprotein containing 

 the 4-chloro-3-azobenzoic acid group. The explanation that we pro- 

 pose of this cross reactivity between one antiserum and the substituted 

 apoprotein, but not between the other antiserum and the different 

 azoprotein, is that the phenomenon depends upon the fact that the 

 chlorine atom is much larger than the hydrogen atom that it replaces, 

 the van der Waals radius of chlorine being about 1.8 A. and that of 

 hydrogen only about 1.2 A. If it is assumed that the combining region 

 of an antibody fits tightly about the haptenic group of the immuniz- 

 ing antigen, the anti-4-chloro-3-azobenzoic acid antibodies would con- 

 tain in the appropriate place a cavity into which a chlorine atom could 

 fit, along with the rest of the haptenic group. This cavity, with radius 

 1.8 A., would be large enough to accept easily a hydrogen atom in the 

 unsubstituted azoprotein, and the replacement of chlorine by hydro- 

 gen would have no effect other than to decrease slightly the force of 



