354 C. E. SCHWERDT 



assays, whether based upon the all-or-noiie or upon the local lesion type of 

 response, are useful in estimating relative virus concentrations but give no 

 clue as to the total mass or numbers of virus particles involved. If a virus can 

 be purified or if the physical particle with which infectivity is associated has a 

 characteristic size and shape readily recognized by electron microscopy, 

 absolute particle counts can be made and correlated with infectious units. 

 Several excellent techniques for direct coimting by electron microscopy have 

 been devised and applied to a number of bacterial, animal, and plant viruses. 

 Although there are a few indirect methods of enumerating virus particles, 

 they would seem to be less desirable, since they require assumptions regarding 

 purity of preparation, particle mass, or the quantitative aspect of some 

 biological activity, such as hemagglutination, which are not always readily 

 verified. 



Almost all viruses studied so far have revealed a particle per infectious 

 unit ratio greater than one, with the exception of bacterial viruses (T series of 

 coliphages), for which the ideal 1 : 1 relationship has been observed. In spite 

 of the failure of most viruses to exhibit the ideal behavior of one particle per 

 uifectious dose, it is the consensus of many that infectivity may be initiated 

 at the host as well as the cellular level by a single viable virus particle. This 

 opinion has been arrived at by inference from statistical analyses applied to 

 titration data of both the quantal and local lesion type of response and is 

 supported, of course, by direct observation in the case of bacterial viruses. 

 There is a dissenting opinion, however, that the probability of infection does 

 not depend upon the chance presence or absence of a single virus particle in an 

 inoculum but rather upon the variable susceptibilities of the individual 

 members of the host species inoculated, particularly in those systems where 

 large numbers of virus particles are required to induce infection. Variations in 

 host suscej)tibilities may well play a role in determining the nature of the 

 dose-response curve for some virus-host systems yet not necessarily vitiate 

 the "one-particle" concept of infection under optimum conditions. 



The usual experimental observation of a virus particle per infectious dose 

 ratio greater than one may have several explanations. It may be due to a 

 general, more or less uniform, resistance of the host animals. It may also 

 reflect the fact that only a portion of the entire population of particles is 

 infective. The fraction which is noninfective may be so by virtue of thermal 

 inactivation or faulty synthesis during replication. There is a further possi- 

 bility, proposed by Isaacs (1957), that the virus particles present a spectrum 

 of infectivity with some particles showing a high probability and others a 

 very low probability of initiating infection. Whatever the reason, the 

 reciprocal of the ratio of particles per infectious unit expresses the probability 

 that a particle will infect under a given set of conditions, assuming that the 

 "one-particle" theory of infection is correct. 



