NATURAL SELECTION 143 



evolution at which the characters which they share are poten- 

 tially present : they are possible variations. 



It might be thought that, though the specialisation of complex Correla- 

 animals may set limits to the range of their variation, yet that 

 the number of possible forms within those limits would increase 

 with their complexity. Some check is put upon the increase by 

 the correlation of various parts. The subject is very imperfectly 

 understood, but what is known enables us to see that the organ- 

 ism is no mere aggregate of units but that a variation in one part 

 involves a variation in several others. There is a system of 

 mutual dependence between organ and organ, which, as our 

 knowledge^grows, is found to extend and have an importance 

 formerly undreamed of. 1 



I now pass on to what I shall call sequence of variation, the Sequence 

 known likelihood that among the offspring of parents in which f on ana ~ 

 a particular variation has occurred some will vary in the same 

 direction, so that if a flower shows a tendency to doubleness 

 some of the seedling plants from it are likely to show the 

 same tendency in a greater degree. 2 This, of course, does not 

 hold, if the plant is crossed with another that has not the same 

 peculiarity : there must be isolation. Given isolation, there is 

 likely to be a second step in the same direction and, if there is 

 selection, i.e. a series of isolations, the process is likely to continue. 



We have now in our consideration of variations discovered 

 two principles which limit the field within which chance holds 

 sway. 



(1) Even in its variations an organism is bound by heredity: 

 it cannot deviate far. 



(2) There is such a thing as sequence of variation under the 

 conditions just explained. 



After what I have said, the question, however unreasonably, Can we 



TTT i r i e f attribute 



is sure to be put : "Was the evolution or the various forms of the evolu- 

 life of mollusc, crustacean, reptile, bird, mammal necessitated tion fall 



... 3 -r T7 - . species to 



by the nature of the primitive protozoa: Was there an in- an in- 



1 See p. 80. herent 



^ Darwin called this continuity of variation : see Animals and Plants under Domestication, 



vol. ii. p. 241, and index under variation. I avoid the term because variation by 



long sudden steps is now called discontinuous variation. 



