THE PHYSICAL PROPERTIES OF INFECTIVE PARTICLES 301 



yellow mosaic virus. At the other extreme is the problem of identifying virus 

 particles in suspensions where the heterogeneity of particle sizes is so great 

 that division into classes is not realistically achievable. For example, the 

 identification of the mammary tumor agent of mice has proven frustratingly 

 uncertain for this reason (Howatson, 1953), and for the additional one that 

 the infectivity assay is tedious and imprecise. 



Perhaps the most clear-cut instance of the use of the electron microscope 

 in the identification of particle class with infectivity is fomid in the investiga- 

 tions on poliovirus (Bachrach and Schwerdt, 1954). Before the investigations 

 by Bachrach and Schwerdt that led to the identification of the virus particle 

 it had been established by methods of filtration and sedimentation that the 

 infective agent was in the size range 10-50 m^Lt. Electron microscopy of 

 partially purified suspensions from infected central nervous system tissue 

 of cotton rats showed, fortunately, that only two classes of particles were 

 present: spheres of about 10 m^u, and of about 35 m/x in diameter. The smaller 

 particle was present also in control material, presumptively pointing to the 

 larger particle as the virus. This could not be assumed, however, since the 

 possibility existed that the smaller-sized particle was the virus and that 

 similar-sized particles also existed in the control material. Separations of 

 the two particle classes were attempted by means of centrifugation in a 

 special separation cell; these experiments were successful in drastically 

 altering the relative numbers of particles in the two size classes. Correlation 

 of particle counts with infectivity showed that the apparent specific infec- 

 tivity remained constant for the larger particle, before and after the separa- 

 tion, and that it changed greatly for the smaller particle. On this numerical 

 evidence it was concluded that the class of particle with a diameter of about 

 35 m/A contained the virus, and subsequent work has abundantly confirmed 

 this. The great sensitivity of the electron microscope in identification studies 

 is indicated by the fact that none of the purified preparations contamed 

 more than 10~^ gm. of virus protein; actually 10~'^ gm. would have been 

 enough for the spray-drop examination. 



2. Sedimentation 



The ultracentrifuge can be used to identify the particle class containing 

 an infective agent with those physical parameters for a particle that are 

 amenable to sedimentation analysis, namely, the ratio of the molecular 

 weight to the shape factor, and the effective density. In assessing the identi- 

 fications so obtained it is convenient to consider two cases: where the sus- 

 pension is nearly physically homogeneous, and where the suspension is 

 quite inhomogeneous. In the former case one can speak of a "characteristic" 

 particle for which the problem is to see whether or not the infective agent 

 has the physical attributes of this particle class. In the latter case, the virus 



