29. DEOXYRIBONUCLEIC ACIDS AS MACROMOLECULES 23 



of being very uncertain. They are perhaps of the right order of magnitude, 

 but it is difficult tc evaluate the extent of the error. 



As to the radius of gyration R 2 , since it is obtained from the evaluation 

 of the initial slope of the extrapolated curve (c//C) c =o as a function of 0, 

 it is clear that the uncertainty is still much greater. We think that it is 

 wise, in the present state of affairs, not to pay too much attention to the 

 numerical values of R" as they have been published so far. 



It is beyond any doubt necessary to find new devices in which 6* will 

 be much smaller than is now possible. However, one should keep in mind 

 that — since large particles scatter the maximum intensity at 9 = — there 

 will ^t ill remain the experimental difficulty of getting rid of the error intro- 

 duced by the eventual presence of dust in the solutions. 



In Tables I and II there is listed a large collection of values of M w as 

 reported by several authors for samples of different origin and different 

 types of preparation. 



At first sight, the striking feature is the large spread of the experimental 

 values of My, , which range from one million to more than thirty million; 

 a difference larger than the experimental error, even in its more pessimistic 

 evaluation. 



A more accurate examination reveals that the values are not scattered 

 at random. More than 40% of the numbers lie between 6 X 10 and 

 8 X 10 6 (Table I) and between 6 X 10 fi and 10 X 10 6 (Table II); more 

 than 30% are between 10 X 10 and 16 X 10 ; very few values are found 

 to be less than 6 X 10 or larger than 16 X 10 . Even if, keeping in mind 

 our suspicion as to the significance of the numerical values, we consider 

 that M w is nothing more than a parameter more or less specific of a DXA 

 sample with no clear relation to the true mean molecular weight, the fact 



remains that these parameters roughly belong to two groups: the 6 X 10 

 to 8 X 10 6 group and the 12 X 10 e to 16 X 10 6 group. 



It is evident from Tables I and II that the origin of the samples cannot 

 be responsible for this grouping: we find that in each group there are DNA 

 samples coming from bacteria as well as from higher organisms. It is, there- 

 fore, necessary to question the procedure of extraction and ask whether, 

 using the same material, a given technique gives reproducible results. 



This is the reason why — before going on — we shall, in the following sec- 

 tion, examine this very important question. 



b. Conditions of Reproducibility of the Dispersion of DNA Particles in Solu- 

 tion 



A great number of observations have been made in this field. Here we 

 shall sum up the main ones. 



19 C. Sadron, J. Pouyet, and R. Vendrely, Nature 179, 263 (1957). 



