42 R. MAEKHAM 



distribution is discontinuous at first, but on standing overnight it becomes 

 more nearly linear (this may be speeded up by judicious mixing). The virus 

 solution is layered at the top, and the tubes are centrifuged. Each component 

 of the solution then moves down the sucrose gradient as a thin layer. The 

 velocity at which these layers sediment is a variable, because several factors 

 are involved. The centrifugal field increases with the radius (about 2-fold), 

 but the viscosity of the fl.uid increases about 6-fold from the top to the 

 bottom and, of course, the density difference between the virus and the 

 solution decreases by a factor of about 0.5. The last two effects tend to slow 

 down the bomidaries as they descend. 



After the required centrifugation time has elapsed, the centrifuge is stopped 

 and the tubes sampled. In some cases, as with turnip yellow mosaic, the 

 zones are clearly visible, but in other cases samples are taken at definite 

 intervals. (If tubes transparent to ultraviolet light are used a concentration 

 of virus of 10 )ug./ml. should easUy be visible by photography.) 



An improvement on the method has been described by Smclair and 

 associates (1957), who used an apparatus (designed by Prof. E. M. Bock) 

 which produces a linear concentration gradient automatically by mixing 

 the contents of two hypodermic syringes; they used the method to separate 

 the two components of the wUd cuciunber mosaic virus. In a case like this, 

 where the two components have rather differing specific volumes, the effects 

 of the density of the sucrose solution on the actual separation is not nearly so 

 marked as when the substances to be separated have similar specific volumes. 



Ethanol and methanol solutions also present the possibilities of construct- 

 ing gradients, but the density differential over acceptable concentrations 

 (0-20 %) is only about 0.025-0.03, as opposed to 0.15 for sucrose. Another 

 possibiHty is the use of inorganic salt solutions, such as solutions of the 

 halides of the alkali metals, where densities of the same order as those obtain- 

 able with sucrose can be obtained but with lower viscosities. Such solutions will 

 not usually precipitate viruses, but they may damage some of the less stable ones. 



The density gradient method has been used with some success for the 

 purification of some of the larger and not very stable viruses, such as the 

 womid tumor virus (Brakke et al., 1954). This method, combined with the use 

 of radioactive isotopes, might well yield some interesting uiformation about 

 these large viruses which are capable of multiplication in insects as well as 

 in plants. A point of particular interest in this regard is whether they contain 

 ribonucleic acid like other plant viruses, or if, like many of the large viruses 

 of animals, they contain deoxyribonucleic acid. 



E. Following the Course of a Purification 

 The course of purification of a plant virus is best followed by activity 

 measurements. These are not altogether reliable, but the first thing which 



