138 C. A. KNIGHT 



Studies of the chemical compositions and structures of strains of TMV 

 have shown that strains possess highly characteristic common features, such 

 as nucleic acids of the same composition, peptide chains terminating with the 

 same amino acid residues, and macromolecular structures possessing the 

 same numbers of subunits. These features will be elaborated in the sections 

 devoted to the protein and nucleic acid components of strains. 



D. Evaluation of Criteria of Strain Relationship 



From the point of view of establishing strain relationship, apphcabihty 

 and specijficity of the tests employed are both of prime importance, but, 

 fundamentally, specificity is clearly paramount. What then are the most 

 specific criteria? As will be described in later sections, mutation of plant 

 viruses seems, on the basis of present information, to result in relatively 

 small changes in the major chemical and physical properties of viruses. Even 

 those variants which differ most widely in biological properties have been 

 found to possess certain apparently identical physical and chemical char- 

 acteristics not shared with imrelated viruses. Therefore, the coincidence of 

 specific major chemical and physical properties would seem to provide the 

 best evidence for strain relationship. Since the immunological reactions are 

 doubtless based on chemical structure, it seems appropriate that immuno- 

 logical tests be regarded second only in specificity to the structural char- 

 acteristics themselves. By comparison, all of the other criteria enumerated 

 may be considered inferior to the physicochemical and immunological 

 standards with respect to specificity, with the possible exception of Holmes' 

 test for response to a genie change in the host. 



Turning to applicability, it can be pointed out that relatively few of the 

 300 or so plant viruses presently known (Smith, 1957) have been obtained in 

 a state permitting precise physicochemical characterization. On the other 

 hand, immunological tests can often (though not always) be successfully per- 

 formed on smaU amounts of crude materials. In proper perspective, then, the 

 physicochemical tests for strain relationship are to be preferred for the ulti- 

 mate in specificity, but from the point of view of practicality, the one grade 

 less specific but widely applicable immunological tests may be for the present 

 more generally useful. 



Another pertinent question is how many of the criteria of strain relation- 

 ship must be satisfied in order to establish that two plant viruses are variant 

 strains belonging in the same group of viruses. It seems to the writer that 

 this question is answered variably by different workers, but that the follow- 

 ing principles may be suggested: (1) In general, greatest reliance shoidd be 

 placed on the outcome of the most specific tests, namely, the physicochemical 

 and immunological tests. (2) The more criteria satisfied, the stronger is the 



