IMMUNOLOGICAL METHODS IN THE STUDY OF VIRUSES 537 



with the unwinding of tlie fibrils to prevent infection. Under such conditions, 

 however, the phage particle could still adsorb further antibody. 



In this discussion no mention has been made of the important findings of 

 Jerne (1956 ; Jerne and Avengo, 1956) in regard to low-grade antibody 

 against phage T4 (tryptophan-dependent strain) present in normal serum, 

 or during the early stages of immmiization. The topic is, however, more 

 relevant to a discussion of the nature of antibody production than in the 

 present context. 



B. Neutralization of Plant Viruses 



Relatively Httle work has been reported on the neutrahzation of infect- 

 ivity of plant viruses since the work of Chester (1934) more than twenty years 

 ago. The reasons for this relative lack of interest are probably multiple. Most 

 fundamental work has been at the biochemical level; infectivity tests have 

 been largely limited to assessing the viabihty of virus treated in various 

 fashions. Sufficient amounts of virus are available to make precipitin reac- 

 tions the method of choice for immmiological work and methods for estimat- 

 ing infectivity quantitatively by the local lesion method are of only limited 

 applicability. 



The relationship between the amount of virus applied and the number of 

 local lesions produced deviates considerably from the linear, but this can be 

 allowed for by the use of appropriate corrections. A more serious difficulty 

 is that the concentration of virus particles necessary to produce a single 

 lesion on a susceptible leaf is of the order 10^-10' per miUiliter; to obtain 

 usable results m quantitative inactivation experiments, concentrations ap- 

 proaching 10^° particles per milliliter must be used (RajDpaport, 1957). In 

 contrast to the type of experiment used in studymg the action of antibody on 

 bacterial or animal viruses, this means that aggregation occurs m all virus- 

 antiserum mixtures and no formal comparison with the results obtained with 

 standard animal and bacterial virus neutralizations is possible. 



A discussion of the older results was given by Burnet et at. (1937). As 

 might be expected from the high concentration of virus antigen used in 

 infectivity tests, the neutralizing action of serum is best showTi when virus 

 is used as highly diluted as is practicable. Mixtures of relatively concen- 

 trated reagents show a high degree of reactivation on dilution. Bawden (1943) 

 quotes a mixture which undiluted gave an average of two lesions. Successive 

 5-fold dilutions of this mixture gave, respectively, 4, 9, 9, and 6 as the 

 average numbers of lesions; the last figure would correspond to 3750 lesions 

 undiluted if a linear relation held. 



More recent experiments by Rappaport and Siegel (1955) and Rappaport 

 (1957) are in general accord. Working with tobacco mosaic virus, their most 



