538 F, M. BURNET 



striking experimental result was the persistence of a considerable fraction 

 (3-6 %) of infectivity in a zone of considerable antibody excess. After 

 removal of aggregates and any uncombined virus by centrifugation, the 

 supernatant fluid contained antibody excess, as tested either in vitro or by 

 inactivation tests. The explanation favored, although it was recognized that 

 others were possible, was that particles with critical sites unoccupied by 

 antibody could be trapped within the aggregates and thus protected from 

 antibody excess. Little use has been made of neutralization tests in the 

 characterization and differentiation of plant viruses, but enough has been 

 done to indicate that the method has potentialities. Spooner and Bawden 

 (1935) showed that antisera against potato virus X failed to neutralize 

 potato virus Y and tobacco mosaic virus. Rappaport et al. (1957), by making 

 use of the shape and position of inactivation curves with a rather wide 

 range of antiserum concentrations, were able to distinguish three serological 

 groups among five strains of tobacco mosaic virus; they consider that the 

 method has important potentiahties. It may be of special use for the char- 

 acterization by serum neutralization of viruses that are difficult to isolate in 

 quantity. 



C. Neutralization of Animal Viruses by Immune Serum 



Despite the extensive everyday use of neutrahzation tests in diagnostic 

 virus laboratories, very httle has been effectively estabUshed about the 

 nature of the process by which infectivity of animal viruses is destroyed. A 

 large part of this difficulty can be ascribed directly to our ignorance of the 

 details of the process by which infection occurs. For every virus that has 

 been studied in detail, with the possible exception of vaccinia virus (Over- 

 man and Tamm, 1956) the ID50, as estimated in the most susceptible avail- 

 able host, corresponds to more than one morphological particle. 



A table has been compiled by Isaacs (1957) to show the approximate 

 number of morjahological particles corresponding to one ID50 for a number 

 of animal viruses. Some examples taken from that table and, where necessary 

 added to as a result of more recent pubhcations are shown in Table I. 



No satisfactory explanation for these discrepancies has been provided. 

 There is no convincing evidence even as to whether in any given case where 

 n particles are needed to produce one initiated infection, the situation is best 

 interpreted (a) as one active particle to w-1 mactive particles, or (b) as n 

 potentially infective particles of which only one will, on the average, initiate 

 a continuing series of infections. Since it is almost invariably found that if 

 two host species are compared a higher proportion of infections are initiated 

 on one than on the other, one has a strong prejudice in favor of alternative 

 (b), but extrapolation of the argument beyond the most susceptible available 

 host may not yet be justifiable. 



