VARIATION IN VIRULENCE 227 



be defined. A strain of influenza virus which, is virulent for man is, in common 

 parlance, either one that spreads rapidly and extensively through the popula- 

 tion, or one that is characteristically associated with a severe morbidity and, 

 for influenza, a relatively high death rate. These effects may obviously be 

 due to widely different and potentially very complex mechanisms. In the same 

 way a strain of poliovirus may be called virulent for man because it paralyzes 

 a high proportion of those it infects, or because it kills a high proportion of 

 those it paralyzes. These vague uses of the term virulence, sanctioned by 

 convention, merely amount to calling any strain virulent if it has high 

 destructive power of the type associated with that virus. There are obviously 

 many grades of virulence, not the simple alternatives we see in bacterial 

 viruses. This usage does emphasize the medically and socially important 

 aspect of the virus diseases of man and his domestic animals, but it deals only 

 with the end product of a multitude of unknown factors, and tells us nothing 

 at all of the possible mechanisms involved. There are more sophisticated 

 definitions, but not one of these can be framed so that it will cover the 

 activity of all animal viruses. 



With this definition of virulence, it is apparent that nearly all examples of 

 adaptation of viruses to new hosts involve increase in virulence for the new 

 host. The field under review therefore covers nearly all examples of alteration 

 in the biological behavior of viruses. Two ways of treating the subject present 

 themselves. One would be to catalogue all examples of variation in host 

 spectrum among animal viruses. This might be useful if the literature were 

 both interesting and as yet inadequately reviewed, but it would be dull, brut- 

 ish, and long. Catalogues of examples of variation in animal viruses have 

 already been prepared (Findlay, 1936, 1939). 



The other procedure, which will in fact be followed, is to try to break down 

 the process of "adaptation" into its component parts, starting with the 

 simplest system, i.e., animal cells in suspension, which is analogous to the 

 bacteria which serve as hosts for bacterial viruses. With this start it will be 

 possible to progress steadily through systems of increasing complexity where 

 adaptedness necessarily constitutes a more complex property, and a strain of 

 virus may be "adapted" or "unadapted" for any one of several widely 

 different reasons. Ultimately, we shall discuss the only available example 

 where the natural evolution of virulence of virus in virgin territory has been 

 studied. First, we shall discuss briefly what is known of the mechanisms of 

 virus variation. 



II. Mechanisms of Virus Variation 



For all living creatures, natural selection copes with, and occasionally 

 profits by, the instability of the genome. So it is, too, for viruses. Survival of 



