130 



THE CELL AND PROTOPLASM 



Kawlins, who noted that double refraction 

 of flow was exhibited by the juic3 from a 

 plant diseased with tobacco-mosaic virus 

 but not by the juice from normal plants. 

 In view of Freundlich's work on the vana- 

 dium pentoxide sols and because of the na- 

 ture of the double refraction of flow, they 

 concluded that tobacco-mosaic virus or some 

 material regularly associated with it was 

 composed of rod-shaped particles. Since 

 that time, and following the isolation of the 

 virus in purified form, an imposing mass 

 of evidence has been obtained, chiefly by 

 Lauffer, that tobacco-mosaic virus consists 

 of molecules having a cross section of about 

 12 mp and a length of the order of 400 

 mp. The strains of tobacco-mosaic virus, 

 the related cucumber-mosaic-3 and -4 vi- 

 ruses, and latent-mosaic-of-potato virus 

 have also been found to have similar rod- 

 like shapes. Tobacco-ring-spot, rabbit- 

 papilloma, vaccinia, and bushy-stunt-of- 

 tomato viruses have not been found to show 

 double refraction of flow and hence prob- 

 ably have relatively symmetrical shapes. 



An interesting phenomenon shown by 

 viruses having a very asymmetric shape and 

 apparently directly dependent upon a rod- 

 like shape, is the formation of two distinct 

 layers when rather concentrated solutions 

 of such viruses are allowed to stand. The 

 line of demarcation is very sharp and 

 gradually rises with time. The upper 

 layer is the more dilute and shows double 

 refraction only when caused to flow, 

 whereas the lower layer is the more con- 

 centrated and is spontaneously doubly re- 

 fracting. The lower layer appears to result 

 when the rod-shaped virus molecules be- 

 come sufficiently concentrated so that they 

 lose their ability to rotate about their two 

 shorter axes, and appears to consist, there- 

 fore of a 3-dimensional mosaic of regions 

 arranged at random to each other but in 

 each of which all of the molecules lie ap- 

 proximately parallel. The phenomenon ap- 

 pears to be quite analogous to that first 

 reported by Zocher and Jacobsohn for 

 vanadium pentoxide sols. Lauffer con- 

 cluded that the lower layer represents a 

 special case of double refraction because 



it has no extinction direction. He found 

 that the double refraction exhibited by to- 

 bacco-mosaic virus is due largely, if not 

 entirely, to the shape of the particles, and 

 scarcely, if at all, to intrinsic double re- 

 fraction. 



The marked asymmetry of some of the 

 viruses has thrown doubt upon sizes esti- 

 mated from ultrafiltration, ultracentrifuga- 

 tion, and diffusion data, since few studies 

 have been made on the manner in which 

 rod-shaped particles filter, sediment, or dif- 

 fuse. Frampton has stressed the facts that 

 certain moderately concentrated solutions 

 of tobacco-mosaic virus do not obey Po- 

 iseuill's and Fick's laws, and that even 

 dilute solutions exhibit anomalous viscosity. 

 He interprets these results as being due to 

 inter-particle attraction and considers that 

 molecular weight values calculated from 

 sedimentation, diffusion, and viscosity data 

 are wholly ambiguous. Although it is 

 generally accepted that in moderately con- 

 centrated solutions inter-particle forces 

 are present which are sufficiently strong to 

 influence the rates of sedimentation, diffu- 

 sion, and viscous flow, there appears to be 

 little justification for assuming such an 

 extreme viewpoint; the data from which 

 molecular weight values were calculated 

 were obtained with dilute solutions, and 

 there is much evidence which indicates that 

 these forces are negligible in dilute solu- 

 tions. Furthermore, Robinson recently 

 showed that even the anomalous viscosity 

 exhibited by dilute solutions may be ex- 

 plained without reference to inter-particle 

 attraction. It may be concluded, therefore, 

 that even in the case of asymmetrical vi- 

 ruses such physico-chemical data may be 

 used with considerable justification in the 

 calculation of molecular weights. In the 

 case of tobacco-mosaic virus, Lauffer has 

 shown that different methods of estimating 

 molecular weight from viscosity, sedimenta- 

 tion, and diffusion data, one of which is 

 independent of assumptions concerning the 

 shape of the particles, give values that 

 agree closely. The cross section of the 

 molecule, calculated from the length to 

 width ratio of 35 : 1 which was obtained 



