EVOLUTION AND VIRUSES 



identify, is in pathogenicity, either in the type and severity 

 of symptom caused or in the host range, but other properties 

 vary no less. Changes in transmissibility have already been 

 mentioned, but resistance to heat or other inactivating treat- 

 ments can vary, as can antigenic and aminoacid constitutions, 

 both in strains isolated experimentally and those found in 

 naturally infected plants. 



Some naturally occurring variants of tobacco mosaic virus 

 differ greatly from one another in constitution and general 

 properties. As these strains were usually found in taxonomic- 

 ally unrelated plants, their differences seemed to entail long 

 periods of evolution in separation from one another. However, 

 a strain from leguminous plants with which I have worked 

 changes its constitution greatly on immediate transfer from 

 leguminous plants to tobacco. In tobacco plants it closely 

 resembles the type strain of tobacco mosaic virus, except in 

 its ability to infect leguminous plants systemically; this it does 

 not do very readily, but when it does it changes to a form 

 that contains aminoacids not present in the tobacco form and 

 that is no more like the type strain than others that are 

 thought to have reached their great differences only after 

 they had evolved for a long time along different lines. The 

 fact that large reversible changes can happen quickly is not 

 evidence that only one genetic factor is involved. The change 

 from the form that invades tobacco to the one that invades 

 beans need happen at only one out of hundreds of infection 

 sites, at each of which many thousands of particles are 

 produced. When one appears with the immense advantage over 

 the others of being able to invade its new host systemically, 

 it will escape from them and multiply unhampered by com- 

 petitors. The appearance is created of the whole stock of virus 

 changing its character, but this is apparent rather than real. 

 Most of the initial infecting particles multiplied unchanged, 

 and to produce the one with the form needed to infect the 

 new host systemically may have involved many separate 

 genetic changes, mutations, recombinations, or both, during 

 the local multiplication of the original inoculum. 



The differences so far noted between the constitution of 

 different strains of viruses lie in the protein fraction, and are 

 detectable by amino-acid analyses or by serological tests. 



23 



