CH. XX] THE ORIGIN OF SPECIES 215 



one can imagine a genus arising from another by successive 

 mutation of a large number of the characters of the latter. 



But if evolution be thus to an appreciable, if not (as Jeems 

 more probable) a very large, extent predetermined and governed 

 m Its unfoldmg largely by definite laws, or by mechanical con- 

 siderations like age, then it is clear that it is no longer safe to 

 consider that advantage to the species has had anything to do 

 with the actual evolution of that species, though" it will have 

 determined to a very large extent whether or not that species 

 shall survive. It maij have been directly concerned in the evolu- 

 tion, but it will be safer to leave it out of consideration, and to 

 study evolution in much more detail before committing our- 

 selves. This study must be especially from an experimental 

 standpoint, perhaps largely Mcndelian, and we must, it seems 

 to me, work without any ulterior idea of any aim to which 

 evolution may be directed (even the very local one of immediate 

 advantage to the species), until we really possess some facts 

 upon which we may reconstruct a theory of its operations. The 

 work described in this book is largely iconoclastic, and I do not 

 propose, in the present volume, to try to substitute any new 

 theory of evolution for that which has for so long held the field, 

 but merely to suggest a point of detail in which the latter theory 

 may in my opinion be altered with advantage, by the acceptance 

 of the theor,y of mutation, whilst in a later work I shall attempt 

 to bring forward some of the conclusions about evolution to 

 which the latest extension of the work upon Age and Area 

 has led. 



If we remove advantage from the list of factors that may be 

 operative in evolution— and it is clear that at most it can only 

 be a small one— then it is evident that the mutations that dis- 

 tinguish species from one another cannot proceed in easy stages, 

 unless there be, as is of course by no means impossible, some at 

 present inscrutable law guiding them. The whole change, it 

 would seem, must take place at once. And this brings us to the 

 question of how large a mutation may be. 



Size of Mutations. Many people think that a mutation must 

 be very small, like the differences in the ".lordauian" species of 

 Erophila verna which are so numerous in Europe, or in the 

 British Ruhi or Hieracia. My own opinion, which I ha\-e held 

 for the last eighteen years, and have published on various 

 occasions (especially in 123, p. 329), is that this is simply placing 

 an unnecessary handicap, for which there is no positi\'e evidence, 



