144 ALICK ISAACS 



are still two possibilities with regard to incomplete virus. It may be thought 

 of as deformed virus, produced as a result of a faulty cycle of multiplication, 

 or it may be immature virus prematurely released from cells before its 

 development had been completed. At the moment, it seems to be logically 

 impossible to distinguish between these two possibilities, and it may be best 

 to regard incomplete virus simply as the product of an abnormal cycle of 

 virus development without specifying the abnormality further. Unfortunately, 

 most of the work on incomplete virus has been carried out with influenza 

 viruses, although recently there have been hints that the same phenomena 

 may apply to other viruses. The formation of incomplete virus is such an 

 interesting phenomenon that search among other viruses seems well worth- 

 while. 



A. Properties of Incomplete Virus 



Incomplete influenza virus, as defined by von Magnus (1946), and prepared 

 by serial passages of undiluted allantoic fluid virus, has a very low infectivity 

 relative to its agglutination titer. In comparison with standard influenza 

 virus, incomplete virus may have a ratio of infectivity/hemagglutinin titer 

 (I/HA ratio) of 10 x to 10 -5 of that expected. The hemagglutinating behavior 

 of incomplete virus does not differ from that of standard virus as determined 

 by the number of virus particles (seen on electron microscopy) per agglutinat- 

 ing dose (Werner and Schlesinger 1954; Donald and Isaacs 1954a). Hence, the 

 characteristic property of incomplete virus is its low infectivity per virus 

 particle. 



In addition to its low infectivity, incomplete virus is also reported to have a 

 low toxicity when injected intracerebrally into mice (Bernkopf, 1950). 

 However, it appears to be a good interfering agent, as shown by its ability 

 after inoculation into the chick embryo to suppress hemagglutinin production 

 by a large dose of standard influenza virus given 18 hours after the incomplete 

 virus (von Magnus, 1954). These findings are much more significant than 

 earlier results reported by von Magnus (1951a) of "autointerference" induced 

 by incomplete virus inoculated intranasally into mice, since the interpretation 

 of autointerference in mice is complicated by the fact that the incomplete 

 virus would provide an antigenic stimulus which might mimic interference. 

 This possibility can be inferred from further experiments by von Magnus 

 (1951a) from which it is clear that incomplete and standard influenza virus 

 showed very similar antigenic behavior, both in in vitro hemagglutination 

 inhibition tests and when formolized and tested for their antigenicity in mice. 

 In the latter case, the antigenicity of the two preparations was proportional 

 to their agglutination titer (i.e., particle count) and not to their infectivity. 

 This finding implies that incomplete and standard virus may have a similar 

 surface structure; further evidence for this came from experiments indicating 



