36 C. L. SADRON 



the limits of experimental error, with the value corresponding to the Watson- 

 Crick double helix (2.10 4 per 100 A.). This linear configuration is not a 

 single straight line (or rod), nor a continuous and flexible coil such as, for 

 instance, in the case of the single-stranded particles of RNA. It could be a 

 large zigzag chain, composed of elements of a length larger than A/20, that 

 is to say larger than about 300 A. The molecular weight of each element 

 should then exceed 00,000. If the molecular weight of the total molecule is 

 about 6 to 8 million, this means that each zigzag is formed by less than a 

 hundred rod elements. According to viscosity and stream double refraction 

 this zigzag chain should be practically rigid. Perhaps a rotation of small 

 amplitude — due to thermal agitation — could occur at the bending points. 



Finally, there could be branching or, less probably, loose bundle of ele- 

 mentary rods. All these possible configurations have been enumerated in 

 Section II and presented schematically in Fig. 4. In the present state of af- 

 fairs, neither optical nor hydrodynamical methods permit us to choose be- 

 tween the different figures. If we want to do this, we must rely on other 

 types of considerations. For instance, we may make the following specula- 

 tions. 



It is evident that, if the DNA particles have the double helix structure, 

 and if their molecular weight is of several million, their total extended 



o 



length should be of several thousand A., that is to say of the order of half 

 a micron. Such a long filament has to be closely packed in the nucleoprotein 

 particles in the chromosomes. This means that the double helix line has to 

 be pleated abruptly — for instance like an accordion — enclosing the protein 



( a ) 



>«n><n><n><ii>siii ii><n><n><n><tiiK 



( b ) 

 Fig. 19. (a) Weak points due to an interruption of one of the streams, (b) Weak 

 points due to a local untwisting of the double helix (Ambrose). The proportions have 

 not been respected (the length of each double-helix segment is larger than 300 A.). 



