300 H. K. SCHACHMAN AND R. C. WILLIAMS 



into classes that are physically distmguishable and in discovering which 

 class is associated with infectivity. The identification is considered good if 

 (1) the methods of separation into classes are precise and the infectivity 

 assays are reliable, and (2) identifications secured by more than one method 

 are in accord in their implied physical descriptions of the particle class. 

 Separation may be achieved in several ways. One method is to disperse an 

 aliquot of the particles for direct, but destructive, observation (electron 

 microscopy); the separation is then made mentally by recording numbers of 

 objects in different morphological classes. The other methods require the 

 virus suspension to be handled in a nondestructive manner, since the material 

 as separated must be assayed in that condition. Separation is accomplished 

 in terms of shape, size, and density (ultracentrifugation); shape, size, and 

 possibly electrical charge (filtration); surface potentials and, perhaps, shape 

 and size (electrophoresis). Only the first of these three latter methods will 

 be discussed here, since it is the most commonly used one. 



1. Electron Microscopy 



Since virus particles are noninfectious subsequent to electron microscopy, 

 one must relate morphologies with infective assay by use of aliquot samples. 

 Some form of representative-field technique is employed in which relative 

 numbers of particles in various classes of shapes and sizes can be counted. 

 The spray-drop method, utilizing polystyrene latex spheres in a known 

 number concentration, is particularly convenient. The relative concentration 

 of particles in different classes is determined and an infectivity assay is made. 

 This yields what may be called the apparent specific infectivity of each class. 

 Another preparation must then be made in a mamier such that there is a 

 considerable change in the relative concentrations of the particles of difi"erent 

 classes. New counts and infectivity measurements provide a new set of 

 apparent specific infectivities. If there is a class of particle for which the 

 apparent specific infectivity remains constant, while it changes for the 

 others, the virus most probably resides in this class. 



While in principle the foregoing account is a prescription for the identifica- 

 tion of virus particles by electron microscopy, in practice there are circum- 

 stances that render identification both simpler and more complex. In the 

 case of several viruses, particularly the bacterial and some of the plant 

 ones, a preparation is likely to appear quite homogeneous in the electron 

 microscope. Unless there is reason to believe from other evidence that the 

 infective units are of an altogether different order of size (as with some of 

 the larger animal viruses), and unless the infectivity assay is negative, the 

 presumption is valid that the monodisperse particles seen are the virus. 

 Examples of this readily achieved state of purity are found in the T-bacterio- 

 phages, TMV, bushy stunt virus (BSV), tobacco ringspot virus, and turnip 



