NATURAL SELECTION 137 



epidermic growths and the differentiation involved no new de- 

 parture, but merely specialisation in two different directions. 

 There may have been sudden advances by not very minute 

 steps ; much increased specialisation of cells, great increase in 

 the size of the hairs or scales. But such advance whether by 

 long or minute steps is only an accentuation of characters already 

 present, and so cannot be said to involve a new departure any 

 more than the variations which I have compared to crystallisa- 

 tion, or to the arrangement of patterns in a kaleidoscope. 1 



Not only does Professor Weldon's formula in many cases 

 break down as a diagnosis of variation, but it does not help to 

 explain adaptations why the right variation appears at the right 

 time. It only aims at proving that there will often or always be 

 a sufficient number of individuals varying in the same direction 

 at the same time plenty of material for Natural Selection, if it 

 is only of the right sort. 



I shall now try to show how the possible amount of deviation 

 from the normal is limited by a principle less mathematical, but, 

 I believe, more efficacious than Professor Weldon's. This prin- 

 ciple, viewed in connection with what I have to say about the 

 nature of the environment will, I believe, help to make clear 

 why it is that the various groups seem, as a rule, to follow each 

 its own line of development instead of wasting their resources 

 on haphazard variations that make no hits and effect no adapta- 

 tions. 



The range of variations possible to a species is limited by Heredity 

 heredity ; the course of its evolution beginning in an age, even llmits the 

 geologically, remote has impressed upon it a certain character, variations 

 and to this character heredity insists that it shall be faithful. 

 Even the lowest organisms that we know, specks of protoplasm 

 in which it is difficult or impossible to find a nucleus, have, 

 no doubt, arisen from simpler forms. They belong to species 

 that have, there is reason to believe, each their own phylogeny, 

 their stages of evolution. Already we have cell-structure ; 



1 In Natural Science, May 1899, I used the term Ntiu Departure to mean a discon- 

 tinuous variation. It is certainly best to keep clear of the word " new" as likely in 

 this connection to mislead. 



