IV. THE CRYSTALLINITY OF VIRUSES 67 



cules and if disturbances, which might be brought about by the decom- 

 position in the protein structure concerning the virus action, was in- 

 significant, then minute particles produced by the decomposition would 

 have the virus action. 



Purified virus particles of both Rattle disease of tobacco and stem- 

 mottle disease of potato were found under the electron microscope to 

 show two peaks in the distribution curves for particle length, one at 

 about 150 mjLi and the other at about 300 m/u (49). This can be 

 interpreted as indicating that in such a case decomposition to half was 

 liable to take place. 



With certatn viruses whose particles are extremely liable to decom- 

 pose into fragments, needle-shaped crystals like those of tobacco-mosaic 

 virus may fail to develop on the addition of ammonium sulphate. In 

 fact, the above mentioned plant viruses of globular particles, such as 

 those of southern bean mosaic and tomato bushy stunt, have never been 

 isolated in needles, but yielded rather in spherical crystals such as 

 rhombic plate, hexagonal prisms, or octahedra. 



The crystal shape may be effected by the degree or the manner 

 of decomposition of the rods. Bawden (25) stated that the causes 

 which determine the crystal form are unknown, and that if inac- 

 tive sap is divided and the two parts purified separately, one some- 

 times crystallized in one form and the second in another, and that if 

 dodecahedral crystals are dissolved and then recrystallized, bipyramids 

 or circular laminae may be produced. According to Markham and 

 Smith (50) when turnip yellow mosaic virus is crystallized from salt 

 octahedra are produced, whereas when crystallized from alcohol long 

 prisms are yielded. 



