INTERFERENCE BETWEEN ANIMAL VIRUSES 179 



in certain host systems to production of noninfectious HA and CF antigen 

 (Schlesinger, 1950, 1953; Cairns, 1951, 1954; Fulton and Isaacs, 1953; 

 Ginsberg, 1954; G. Henle et al., 1955); (3) the demonstration that small 

 noninfectious HA particles can be isolated from the infected CAM even under 

 conditions where the liberated yield consists of "standard" virus: these 

 small tissue-bound particles differ in physical and biological properties from 

 incomplete virus obtained on serial undiluted passage (Granoff, 1955b) and 

 are probably truly representative of a developmental phase of the virus 

 (Henle et al., 1956); (4) incomplete virus may contribute genetic markers to 

 recombinants produced in de-embryonated eggs doubly infected with active 

 virus (Burnet et al., 1954). 



It is known that even standard preparations of virus contain varying 

 amounts of noninfectious HA (Gard et al., 1952). As suggested previously 

 (Schlesinger, 1953), production or accumulation of noninfectious virus may 

 then be expected to occur under conditions favoring abnormal or functionally 

 deficient particles which are present to begin with in a heterogeneous viral 

 seed. We face the dilemma that our definition of infectiousness of influenza 

 virus is synonymous with demonstrable production of new extracellular 

 virus. "If the basis for this idea were applied to bacterial viruses, a temperate 

 particle would be labelled 'incomplete' because it fails to give rise, in the 

 majority of infected cells, to (extracellular) progeny" (Schlesinger, 1953). 

 This viewpoint has been expanded by Burnet (1955, p. 141): ". . . every 

 bacterial population contains an enormous variety of genetically incomplete 

 forms, the great majority of which, being at a disadvantage for survival, give 

 rise to no descendants. The situation is probably analogous for the replication 

 of all viruses. In addition to the standard form, we have a variety of genetically 

 imperfect forms. Only in the case of viruses like influenza, where we have a 

 means of titrating the whole population (the HA titer), is it possible to 

 provide an estimate of the proportions showing various grades of imperfec- 

 tion . . .".If this postulated intrinsic heterogeneity of viral populations is 

 at the root of incomplete virus production, then interference is related to 

 it only insofar as the genetically imperfect forms have a quantitative advan- 

 tage over the standard forms. 



5. Interfering Capacity of "Incomplete" Virus 



Aside from the evidence derived from undiluted passage series in eggs (von 

 Magnus, 1951a,b, 1952, 1954; Fazekas de St. Groth and Graham, 1954), 

 other independent studies have shown that "incomplete" virus can interfere 

 with infectious homologous or heterologous influenza virus. Burnet et al. 

 (1954) have shown this in experiments in de-embryonated eggs. Powell and 

 Pollard (1956) have reported that the interfering capacities of incomplete 

 and artificially inactivated virus are equivalent. Suitable quantitative 



