54 DEOXYPENTOSE NUCLEOPROTEINS AND PROSTHETIC GROUPS 



facts; (b) an impulse to further and crucial experimentation. 

 Therefore, before this structural hypothesis is elevated to the 

 dignity of a natural law, a few remarks may be in order. The 

 model provides a satisfactory explanation, though it probably is 

 not the only possible one, of the unity relationships observed ex- 

 perimentally. Whether, in its detailed form, it does more than 

 describe the structure of that portion of the processed nucleic 

 acid preparation from which the diffraction patterns were ob- 

 tained, remains, however, to be estabhshed. More work on genuine 

 nucleoproteins, especially on compounds with proteins other than 

 histone or protamine, e.g., of the type found in tubercle bacilli^, 

 is urgently required. The existence of two discrete and com- 

 plementary chains has not been demonstrated; no deoxypen- 

 tose nucleic acid preparation has shown noticeable deviations 

 from the 1 : 1 ratios, not even for the ratio of purines to pyrimi- 

 dines. 



It is, however, the second part of the hypothesis, namely, the 

 suggestion of a copying mechanism for the mutual reproduction 

 of the two half-chains, that is more difficult to reconcile with 

 some of what is known. It is assumed that a specific pairing is 

 brought about by hydrogen-bond formation between the nitrog- 

 enous bases of the two chains in such a manner that a base 

 having a keto group at the 6 position (guanine, thymine) can 

 bond only with a base carrying a 6-amino group (adenine, 

 cytosine), and that a purine always bonds with a pyrimidine, thus 

 maintaining a constant distance between the two chains. It will 

 be observed that the model is quite strict as regards certain 

 restrictions governing the size of the partners (purine is linked 

 to pyrimidine) and the position of the hydrogen bonds (purine- 1 

 to pyrimidine- 1 ; purine-6 to pyrimidine-6), but does not require 

 the absence or presence of substituents in other positions. Thus, 

 5-methylcytosine or 5-hydroxymethylcytosine should be able to 

 take the place of cytosine, and uracil and 5-methyluracil (thy- 

 mine) should be similarly exchangeable, both in their positions on 

 the ''positive" or "negative" or during the automatic replication 

 of the chains. That the latter replacement does not occur despite 



