THE DIRECT VIRUSES 377 



polyhedra form sparsely over the virogenic masses; these polyhedra vary in 

 size but may be under 0.5 ju in diameter. They are best observed in relation to 

 the cytoplasm and virogenic masses in HCl-Giemsa preparations. Further 

 development proceeds by the continuing growth of the virogenic masses and 

 by the increase in the number and size of the polyhedra at their surfaces. 

 Polyhedron formation does not take place equally at all parts of these surfaces; 

 some parts may be packed with well-developed polyhedra 0.5-1 /jl in diameter, 

 other parts may be completely free of them. During this stage of the disease, 

 large pores develop within the virogenic masses and polyhedra also arise and 

 develop within these pores. In the fully matured colonies of the moribund 

 cell the polyhedra reach a size of 1.5 fi in diameter. 



c. Different Shapes and Sizes of Cytoplasmic Polyhedra. On the whole, the 

 cytoplasmic polyhedra resemble those of the nuclear diseases but there are 

 differences. There seems to be a greater range of size of cytoplasmic polyhedra 

 in an individual smear, from very small to very large indeed, sometimes as 

 much as 10 /x. This was particularly the case in a cytoplasmic disease of 

 Ourapteryx sambucaria. There is also a tendency on the part of the large 

 polyhedra to lose their many-sided character and to appear almost spherical. 

 In the spruce budworm, the polyhedra are chiefly cuboidal and triangular 

 forms have not been observed (Bird and Whalen, 1954). 



d. Dispersal and Arrangement of Virus Particles inside Polyhedra. It is the 

 writer's opinion, after long experience of cutting thin sections of both types of 

 polyhedra, that the cytoplasmic type is considerably harder and it is more 

 difficult to get good sections without much compression by the knife. This 

 gives the impression that the viruses are somewhat elongated but in fact this 

 is only an artifact due to compression. Furthermore, some of the cytoplasmic 

 viruses are very small, one affecting the larva of Sphinx lingustri measures 

 12m/x, about half the size of the tomato bushy stunt virus (Fig. 19), and it is 

 difficult to observe them at all in sections. Indeed, from observation of the 

 sections alone it would appear that little or no virus was present, whereas in 

 actual fact the virus content is very high. Observation is rendered more 

 difficult by the lack of contrast between the virus particle and the surround- 

 ing polyhedral protein (Fig. 10). 



On the whole, it appears that the particles are arranged haphazardly 

 within the polyhedral crystal. Occasionally, however, sections may reveal 

 what appears to be a regularity of assembly within the crystal (Fig. 

 11). 



e. Reaction of Cytoplasmic Polyhedra to Alkalies. It is in their response to 

 treatment with weak sodium carbonate that the two types of polyhedra 

 differ most. We have seen that the nuclear polyhedra dissolve completely, 

 leaving behind a membrane in which the virus rods are contained (Fig. 12). 

 With similar treatment the cytoplasmic polyhedra dissolve only partially and 



