GENETIC INTERACTIONS BETWEEN ANIMAL VIRUSES 291 



In some recent experiments (Lind and Burnet, 1957b), a recombinant 

 NWSE of high but not completely regular pathogenicity for adult mice by 

 the intracerebral route was used in crosses with MEL. Eighteen pure clones 

 with M -\- in vitro characteristics gave almost uniform findings in regard to 

 neuropathogenicity. In 4-5 weeks mice, inoculation of 1 AD intracerebrally 

 gave only an occasional death but most mice appeared sick on the sixth or 

 seventh day and virus could be isolated from the brain of those killed at that 

 time. All were lethal to mfant mice inoculated 2-4 days after birth. This 

 type nM therefore appeared to represent a rather standard level of incomplete 

 expression of neuropathogenicity. On back-crossing nM with WS — , the 

 WSE type recombinants showed the same virtually stable low level of neuro- 

 pathogenicity. 



The interpretation of these results can hardly go further than an acceptance 

 of the view that the fully neuropathogenic strains are the result of a series of 

 genetic changes and that when the quality is transferred to other types a 

 variety of relatively unstable forms emerge. 



2. Redistribution of Mouse Lung Virulence (MLV) 



In the system we have used in most work, MEL, WSE, and the recom- 

 binants M + , WS — , the stability of pathogenicity for the mouse lung and its 

 consistent linkage with serological type has been most striking. Exceptions 

 are so rare that one tends to regard them as due to contamination or technical 

 error. All strains serologically MEL give lung lesions averaging 0.5-2.5 and 

 never lethal after a dose at the 1 AD level; all strains serologically WS kill 

 mice with complete consolidation at the 1 AD level and usually at one-tenth 

 of that dose. This holds also when heat-inactivated components are used 

 (Burnet and Lind, 1952, 1954a; Lind and Burnet, 1953, 1957c). 



In these investigations, the interaction was between highly virulent WS 

 strains and weakly virulent MEL strains. In an unpublished series of experi- 

 ments a variant of WS — was obtained which was completely avirulent for 

 the mouse lung. This allowed a cross between weakly virulent M + and 

 avirulent WS — strains. The MEL recombinants still retained their standard 

 weak virulence. 



Mouse lung virulence is, however, by no means as stable in other systems. 

 It is consistent with our experience to say that where linkage of mouse lung 

 virulence with serological character is not evident one observes redistribution 

 of virulence. The most interesting example concerns WSE which, as noted 

 above, shows no evidence of instability of its high MLV (mouse lung virulence) 

 in interactions with MEL. When crossed with CAM (no MLV) a proportion of 

 recombinant CAM types showed moderate MLV; among WS — and WSE 

 (isolated from the primary fluid) types there was a wide range of MLV 

 (Burnet and Lind, 1955). Similar variability of MLV was observed in 



