INTERFERENCE BETWEEN ANIMAL VIRUSES 169 



viruses in the allantoic cavity and leading to production of "incomplete" 

 virus will be discussed separately (Section III, B, 4). 



Superficial analogies have led some authors to postulate that other types of 

 autointerference are likewise associated with "incomplete" virus production. 

 For example, Mims (1956) found that serum of mice inoculated with undiluted 

 RVF virus often contained virus of low infectious titer but of antibody- 

 combining capacity equivalent to that of high-titer serum. Quite apart from 

 reservations one may have about the relation of interference to the genesis of 

 "incomplete" influenza virus (see Section III, B, 4), it seems noteworthy 

 that autointerference with other systems is characteristic especially of viruses 

 which are not fully adapted to the experimental host or which are mixtures of 

 particles with different properties (in Mims' experiments, e.g., viscerotropic 

 and neurotropic RVF particles). The inoculation of such viral populations 

 could be akin to that of artificial mixtures of avirulent and virulent particles 

 resulting in (a) suppression of multiplication of the virulent component and 

 (b) protection of the host. To determine the presence of avirulent rather than 

 "incomplete" progeny would then require an indicator host susceptible to the 

 avirulent component. 



This approach has been possible in an investigation of "autointerference" 

 with a strain of Type 2 dengue virus in adult mice (Schlesinger and Frankel, 

 1952a,b; Schlesinger et al., 1958). This strain was adapted to suckling mice by 

 intracerebral passages, and after 4-6 passages its lethal titer became fixed 

 around 10~ 7 . Additional serial passages in sucklings brought about a gradual 

 increase in virulence for young adult mice. At low passage levels, titration of 

 the virus by intracerebral inoculation into adults often resulted in auto- 

 interference, i.e., highest mortality rates among those inoculated with high 

 dilutions of virus. Transfer back to sucklings, as indicator hosts for virus 

 lacking adult virulence, revealed that such virus multiplied in the brains of 

 adult mice even when it produced no illness in them. Had such an indicator 

 host not been available, one might have thought that "incomplete" virus was 

 involved. 



Moreover, after some 25 passages in sucklings, the virus attained the same 

 lethal titer for adults as it had for sucklings. Titrations of such high passage 

 virus never gave a pattern of autointerference. When adult mice were doubly 

 infected with low- and high-passage virus, protection against the latter was 

 observed. The relative effectiveness of protection was proportional to the 

 timing or dosage advantage of the avirulent component. There was, however, 

 one significant exception: Mixtures containing sublethal (0.1 to 0.01 LD 50 ) 

 amounts of the virulent and 1 to 10 3 LD 50 (titer in suckling mice) of the 

 avirulent virus induced mortality rates in excess of the combined rates for 

 the two components given singly. How could this paradox be explained? 

 By analogy to the experiments on reactivation of partially inactivated 



