70 E. MARKHAM 



The virus, in fact, in the dry state, as "dry gel," has a weak positive 

 birefringence, while the virus protein has weakly negative birefringence. 

 The nucleic acid must therefore contribute to the birefringence in a positive 

 sense, probably because the purine and pyrimidine rings are oriented parallel 

 to the axis of the rods. This is in keeping with the ultraviolet dichroism of 

 the particles (Franklin, 1955b). 



The infrared dichroism of the virus has also been investigated and indicates 

 that the polypeptide chains are in the a-configuration, and run at right 

 angles to the long axis of the virus particles (Fraser, 1952; Beer, 1958). It 

 will be noted that this does not define the direction of the chains absolutely. 



VI. Properties of the Virus Protein 



It is usually assumed that the suspensions of the tobacco mosaic virus 

 are uniform both in nature and composition. A very superficial survey of 

 the published work on this virus will indicate that this is not altogether so. 

 For example, it is quite easy to show that any tobacco mosaic virus prepara- 

 tion contains an appreciable number of mutants. It is quite simple, by taking 

 single lesion isolates from Nicotiana glutinosa, to demonstrate one or more 

 "yeUow" variants per hundred isolates, and yeUow variants are not the only 

 ones present. On a more chemical level, simple treatment of the virus 

 suspension at 0°C. at a pH of 10 results in the disintegration of 70 % of the 

 particles, but leaves 30 % of the particles intact apparently indefinitely 

 (Schramm et al., 1955b; Harrington and Schachman, 1956). Moreover, the 

 progeny of such alkali-resistant particles has a similar proportion of labile 

 and resistant particles, the exact proportions possibly varying with the 

 "strain" of virus examined. 



On an even more sophisticated level, the amino acid analyses made on 

 purified viruses tend to show deviations from the simple proportions antici- 

 pated, which are greater than one might expect on the basis of a model virus 

 having uniform particles with a regular simple subunit. And, as has been 

 mentioned, small amounts of plant contaminants are invariably present. 

 It is, therefore, not surprising that some of the more critical workers regard 

 the detailed analysis of viruses as probably misleading. The alternative 

 is to hope that the structure may be uniform, at least to a major extent, but 

 to bear in mind that there may be pitfalls. This is what we shall do, so that 

 it is necessary to bear in mind that the following discussion refers to an 

 ideahzed situation. This is, of course, the only way in which one can tackle 

 this type of problem at the moment, and it is pertinent to remark here that 

 much of the methodology in current use is suited to this type of approach. 

 For example, both ultracentrifugation and X-ray diffraction analysis are 

 techniques which automatically sample the greater part of any material 



