130 ALICE ISAACS 



other types of condition under which hemagglutinin of low relative 

 infectivity is produced, but in both cases the hemagglutinin remains 

 intracellular and is not normally liberated from the cells. The noninfective 

 hemagglutinin produced in these other conditions shows some significant 

 differences from the incomplete virus of von Magnus and therefore will not 

 be called incomplete virus here. The first example was described by Granoff 

 and associates (1950). They prepared extracts of chorioallantoic membranes 

 infected with NDV and found that after high-speed centrifugation (20,000 

 r.p.m. for 20 minutes) the supernatant showed serologically specific hemag- 

 glutinin of low relative infectivity. Virus from the allantoic fluid did not 

 behave in this way after high-speed centrifugation. Granoff (1955) later 

 confirmed these results with NDV and found that the same phenomenon 

 could be demonstrated with the PR8 strain of influenza virus A. He called 

 this hemagglutinin "S" for small, although of course its size was not defined 

 by the centrifugation results, and it might be merely less dense than the 

 elementary bodies. By passage experiments it was shown that the "S" 

 influenzal hemagglutinin did not reproduce incomplete virus on passage, thus 

 differentiating it from the incomplete virus prepared by the technique of von 

 Magnus. Schafer and Munk (1952) have also reported finding hemagglutinin 

 of low relative infectivity in the supernatant after high-speed centrifugation 

 of membranes infected with fowl plague virus, and Schafer (1957) has shown 

 the appearance of this hemagglutinin on electron microscopy (Schafer calls 

 these incomplete forms). This membrane hemagglutinin consists of balloon- 

 like particles of variable size measuring 50-550 nu/, in diameter in the flattened 

 state. It seems, therefore, that in NDV, fowl plague, and influenza infection 

 there develops in infected membranes hemagglutinin of low relative 

 infectivity, sedimenting less readily than the virus elementary bodies; this 

 hemagglutinin is not normally secreted into the allantoic fluid. By contrast, 

 the same membrane extracts also contain hemagglutinin which shows the 

 infectivity and sedimentation behavior characteristic of mature elementary 

 bodies; and this is normally secreted into the allantoic fluid. It seems 

 possible that the so-called "S" hemagglutinin is one of the virus building 

 blocks; further studies of its appearance in the tissues relative to that of 

 mature elementary bodies might help to decide this point. 



The second situation in which hemagglutinin of low relative infectivity is 

 produced in cells is when there is a partial cycle of influenza virus multi- 

 plication, as was found in the mouse brain (Schlesinger, 1950), the chick 

 chorion (Fulton and Isaacs, 1953), and HeLa cells (Henle et al., 1955). In 

 these three sites, infection by strains of influenza virus which have not been 

 specially adapted to these particular cells results in the formation of soluble 

 antigen and hemagglutinin in readily detectable amounts, but fully infective 

 virus is formed to only an insignificant degree. In these three sites, too, the 



