VARIATION AND ITS CHEMICAL CORRELATES 133 



vectors at higlier-tlian-normal temperatures, but below inactivating tem- 

 peratui'es. Here again the environment may be selecting a strain already 

 present or it may be favoring certain mutations followed by selection. 



5. Isolation from Local Lesions 



Several plant viruses, notably those transmitted mechanically, produce on 

 certain hosts infections characterized by discrete local lesions. Each of these 

 lesions appears to have arisen from a single infectious unit, since usually, 

 although not infallibly, only a single strain can be recovered from such a 

 lesion (Kunkel, 1934; Lauffer and Price, 1945; Benda, 1956). Hence, a 

 method for securing new strains is to rub suitable dilutions of infectious juice 

 on leaves of a host which gives the local response; individual lesions are cut 

 out, ground in a mortar in a drop or two of buffer, and inoculated to one or 

 more hosts suitable for distinguishing new strains. 



C. Criteria of Strain Relationship 



Quite often, plant viruses are found in nature whose relationship, if any, 

 to known viruses is obscure. The question then arises: is this a hitherto un- 

 described agent, or is it a strain of some previously reported virus? Likewise, 

 when two or more distinguishable viruses are found in the same plant, the 

 problem is whether these are strains of the same virus or unrelated viruses, 

 for both situations appear to occur (Kmikel, 1944; Bawden, 1950; Bennett, 

 1953). Fortunately, two early observations provided a rational basis for an 

 approach to these problems. 



One of these was the finding that tobacco mosaic virus is antigenic and 

 that closely related viruses (that is, viruses which on the basis of other 

 properties are suspected of being strains) display serological cross-reactions 

 (Purdy, 1928, 1929; Kunkel, 1947). The other significant observation was 

 the demonstration of cross-protection from infection by closely related 

 viruses (Thung, 1931; Kunkel, 1934). Thus, the first two criteria of strain 

 relationship among plant viruses were immunological and biological in 

 nature. These tests have been very useful over the years and are still widely 

 employed. However, as studies on strains have progressed, additional 

 criteria have developed. An attempt to coUate the possible standards for 

 strain relationship among plant viruses has been made (Knight, 1955a), 

 yielding (with shght modification in the fisting) the following criteria: 



1. Positive serological cross reactions 



2. Positive cross-protection tests 



3. Similarity in host range 



4. Similarity in method of transmission 



5. Similarity in response to genetic change in host 



