218 LINUS PAULING 



helix, to form the extended chain A — A-i, with length twice as long. This would 

 then serve to produce a compound helix 



A— A-i 



A-i— A ' 



with the same total length. This long double helix would then have to be un- 

 twined, and each of the halves A — A~i and A"i — A would coil to form the 

 original complex and its duphcate. There would thus be two processes of un- 

 twining involved, rather than a single one. If the second process of untwining 

 did not occur, we might be satisfied to accept the long double chain as a complex 

 of a molecule A — A-^ and its self-complementary duphcate, A~^ — A. This 

 process cannot, however, be accepted, because there is no mechanism for en- 

 suring the retention of complementariness of one end of the chain and the other 

 end. Mutation might occur, changing the nature of one end or the other end, and 

 if the chain in the process of duphcation did not fold at some stage into the 

 configuration of the short double helix, the lack of complementariness of the two 

 ends of the chain would not interfere at all with the process of duplication, and 

 this property of self-complementariness would be lost. 



There is a significant difference between the manufacture of antibodies and the 

 duphcation of genes. The antiserum that is produced in response to the injection 

 of an antigen is highly heterogeneous (the heterogeneity may correspond to the 

 range of a factor of ten thousand in the equiHbrium constant for combination of 

 antibody and homologous antigen) [12], whereas the duphcation of genes seems 

 to be perfect, or nearly perfect. The manufacture of an antibody molecule is a 

 process that occurs only once, so far as that molecule is concerned, and the mole- 

 cule is not itself then used as a template. If the conditions within the cell in which 

 the antibody molecule is being manufactured happen to favour the formation of 

 a good molecule with large complementariness to the antigen molecule, the com- 

 bining constant will have a large value, whereas if the conditions are unfavourable 

 the combining constant for the antibody molecule that is formed will have a 

 small value. There is no process of trial and error, such that the system can 

 improve in the manufacture of antibodies against the injected antigen. The 

 process of duphcation of genes is different. A gene that had undergone mutation 

 — for example, that has been damaged as the result of the absorption of a quan- 

 tum of radiation — may serve as a template for the manufacture of a complementary 

 molecule; we may call the mutated gene Ai ^ The complementary^ molecule 

 will then serve as a template, in the next cell division, for the manufacture of a 

 molecule complementary to it, which we may call Ao. In general we would not 

 expect A2 to be an exact duphcate of Ai. The process of two-stage duphcation 

 would then continue: A2 ^ A3, A^-^, A4, Af ',.... If after any number of 

 duplications of this sort a molecule, A„, happens to be formed that is especially 

 well suited to serving as a template for the manufacture of a duplicate of itself, 

 the process of duplication from then on would be a rehablc one. The mechanism 

 of heredity thus permits the ultimate discovery of a mutated gene that undergoes 

 true duphcation; but the mutated gene may, and in general will, not be the 

 original gene that has been damaged by X-radiation or produced in some other 



