VIRUS PARTICLES AND THEIR FUNCTIONAL ACTIVITY 337 



The sensitivity of the plaque assay of animal viruses varies greatly 

 depending upon the virus and cell system used. For example, at one 

 extreme, Granoff (1955) found that plaque titers of PR8 and MEL strains of 

 influenza A assayed on epithelial cell cultures from chick embryo lungs were 

 800-fold less than the 50 % egg infectious titer. In contrast, Dulbecco and 

 Vogt (1954a) fomid that titers of poliovirus, type 1, obtained by plaque assay 

 on monkey kidney cultures were slightly higher than the infectivity titer 

 obtained by monkey titration. In studies relating physical particle to infective 

 particle counts, it is desirable, obviously, to work with assay systems of the 

 highest possible efficiency. Such needs have stimulated the search for more 

 susceptible cell lines which can be rewarding, as exemplified by Fogh and 

 Lund's (1955) finding that plaque titers of poliovirus types 1 and 2 were 3 to 6 

 times higher on monolayer cultures of human amnion cells than on monkey 

 kidney cell cultures. 



As has been found witli bacterial virus plaque assays, a linear relationship 

 exists between plaque count and relative virus concentration for most of the 

 animal viruses assayed by this method thus far. On the basis of this propor- 

 tionality, Dulbecco and Vogt (1954a) have established on statistical grounds 

 that a single virus particle is sufiicient to produce a plaque (Section IV, A). 

 Thus, we have available to us not only an accurate method of measuring 

 infectivity but also a means of isolating pure lines of virus from single clones. 



The reproducibility of plaque titers is similar to that noted earlier for 

 bacterial virus plaque assays (Section II, B, 2). For a plaque count of 100 on a 

 single plate, the estimated standard deviation is 10, assuming that the number 

 of plaques reflects a Poissonian distribution of the virus particles in suspension 

 in the inoculum. Thus, upon repeated titrations, the highest count should not 

 exceed the lowest by more than 50 % for a deviation from the mean of not 

 more than 2 standard errors (Dulbecco, 1955). In contrast, a 50 % end-point 

 assay (e.g., poHovirus in cotton rats), in which serial 3-fold dilutions are 

 inoculated into groups of 6 animals per dilution, wiU yield a SE of approxi- 

 mately 0.2 log units with an estimated variation of 600 % between highest and 

 lowest ID50 expected within the two SE deviations about the measured ID50 

 (Schwerdt and Merrell, 1952). 



C. Factors Affecting Virus Infectivity Titrations 



In most cases, every virus particle in suspension is not capable of infecting 

 a susceptible host; hence an infectivity titer is a relative measure of virus 

 concentration and usually represents some unknown fraction of the total 

 number of virus particles present. Various environmental and host factors 

 may have a marked effect upon the efficiency and reproducibility of assay. It 

 is necessary, therefore, to control the conditions of assay well if meaningful 



