44 EXPERIMENTAL EVOLUTION LECT. 



passages concerning this matter might be quoted from 

 his Histoire naturelle generate des Regnes orga- 

 niques, and his Influence du Monde ambiant, etc. One 

 will be enough, " Since Nature," he says," left to 

 herself never allows us to witness considerable modifi- 

 cations in the conditions of life, it is clear that only 

 one way is open to us if we wish to perceive such 

 modifications and to examine their effects on the 

 organism ; we must oblige Nature to perform that 

 which she would not spontaneously accomplish." 

 (Hist. Nat. Gen. iii. p. 389.) 



This is exactly what we require. While facts of 

 observation are sufficiently numerous to give us a fair 

 idea of the amount of natural variability and variation 

 although much may yet be done to give an adequate 

 notion of the amount of this variability we require 

 to extend our knowledge concerning the causes of vari- 

 ability (the natural causes, of course), and to discover 

 in what manner, and to what extent they do operate. 

 We are already acquainted with some of these causes, 

 and we know that by selection, crossings, modified en- 

 vironment, much has been done. But still more can be 

 done, and in experimental transformism lies the only 

 test which we can apply to the evolutionary theory. We 

 must use all the methods we are acquainted with, and 

 also those, yet unknown, which cannot fail to disclose 

 themselves when we begin a thorough investigation 



