454 H. FRAENKEL-CONRAT 



of tobacco known to respond differently to the two parent strains of virus 

 that may have become genetically mixed in the given experiment. If evidence 

 is thus obtained that a given lesion contains genetic material not correspond- 

 ing to one or the other parent strain in any one respect, then this material 

 is further cycled through the same hosts, thus passing again through a single 

 local lesion, to ascertain whether any difference from the parent strains has 

 become genetically fixed. Only if this aim is achieved is the infectious agent 

 regarded as a mutant, and only if the same type of change is observed in the 

 progeny of an appreciable fraction of the first crop of local lesions (i.e., 30 %), 

 does it appear justifiable to assume that the observed mutation is the result 

 of the in vitro mixing of two parental components. 



As stated, the mixing of protein and nucleic acid from different strains has 

 not resulted in such mutants. The seemingly more promising approach was 

 then taken of mixing two nucleic acids derived from different strains, and 

 then reconstituting virus particles by the addition of one or the other protein 

 and buffer. Evidence that a mixing of two types of nucleic acid may actually 

 occur at low ionic strength has been presented in the preceding section. 

 Most of these experiments were performed with common TMV and the HR 

 strain, because the marked biological and chemical differences between those 

 two would have facilitated recognition of an intermediate character. In this 

 case, the first local lesion test already supplied quantitative evidence for the 

 relative amounts of the two viral types in a given solution. For, as previously 

 stated, HR, in contrast to all other TMV strains, evokes smaller lesions on 

 N. glutinosa which do not spread appreciably with time and are thus clearly 

 distinguishable about one week after inoculation. It appeared from such 

 experiments that HR-type lesions could be obtained only if the mixture of 

 nucleic acids used for reconstitution together with TMV protein contained 

 appreciably more HR-RNA than TMV-RNA (preferably 9:1). It further- 

 more appeared that every nonspreading (i.e., HR-type) lesion contained virus 

 resembling HR in aU respects upon further transfers, while the other lesions 

 contained Tj\IV-like material. In the rare instances when mixed symptoms 

 were observed upon transfer, these generally disappeared upon further pro- 

 pagation. The conclusion from a considerable amount of work of this type 

 was that no predictable mutant could be produced by mixing the nucleic 

 acids of HR and common TMV. This failure to obtain genetic evidence for 

 mixing might be interpreted as evidence that each infectmg particle was 

 purely of one or the other strain type, and that the postulated mixed virus 

 rods were actually noninfective. However, since the observed symptoms are 

 the outcome of countless replication cycles, the interpretation appears 

 equally probable that, depending on the proportion of the two components 

 in a given infectious particle, one or the other type of progeny becomes 

 dominant and leads to exclusion of the other. 



