THE BIOCHEMISTRY OF PLANT VIRUSES 41 



from the axis of rotation, the regions between the boundaries in a cell or 

 tube are in a constant state of mixing. This is, of course, the reaction in 

 the fluid to the centrifugation, which tends to cause a decrease in concentra- 

 tion with increasing radius which has to be eliminated by convection. In 

 addition to this inevitable stirring, there is a tendency for the periphery of 

 centrifuges to heat more by air friction. This causes a decrease in density of 

 the fluid at the bottom of the tubes, and so also gives rise to stirring. Because 

 of these effects, and because of vibrations due to imperfect balancing, the 

 boundaries tend to be unstable. In a good ultracentrifuge stability is achieved 

 down to very low concentrations, but in ordinary angle centrifuges running 

 in air, boundaries cannot remain stable if the concentration is less than a few 

 milligrams per milliliter for substances having sedimentation coefficients of 

 less than 100 S. When this happens the rate of sedimentation drops off 

 rapidly. 



When this takes place most of the slower component remains in the 

 supernatant fluid suspended by the convection in the centrifuge tube in 

 much the same way as a piece of thistledown is suspended m a breeze. The 

 more rapidly sedimenting component is not so much affected by the stirring 

 especially if it is present in greater amount at the start. The pellet which 

 forms is almost pure, but the supernatant fluid is, of course, a mixture. 



The more slowly sedimentmg component may be isolated in a reasonably 

 pure condition by spimiing the mixture in sufficient concentration for two 

 boundaries to form (Fig. 2). The upper layer of material may then be sampled 

 by means of a pipette with a fine tip. This method is ideal for making material 

 of 80 to 90 % purity, but 100 % purity cannot be realized; the only way to 

 attain this is by an ultracentrifuge cell having a filter paper layer at the 

 bottom (Markham, 1953b) or by means of the method of gradient centri- 

 fugation. 



2. Gradient Centrifugation 



A technique which has great potentialities in the separation of unstable 

 viruses and for the separation of mixtures, such as the two components 

 of turnip yellow mosaic or wild cucumber mosaic viruses (q.v.), is the density 

 gradient centrifugation method developed by Brakke (1953, 1956). 



The idea of using artificial density gradients to stabihze centrifugation 

 and electrophoresis is one which has occupied the minds of workers for a 

 long time, but the full potentialities of this method were not realized imtil 

 the S.W. bucket rotors were developed by Spinco. Brakke took advantage 

 of these rotors for his experiments, and he proceeded roughly as follows: In 

 each tube equal volumes of 40, 35, 30, down to 5 % sucrose or an equivalent 

 series in buffer solution are layered successively. One then obtains a variation 

 in density from 1.02 at the top of the tube to 1.17 at the bottom. The density 



