PRIMARY OR ORIGINATIVE FACTORS. 767 



The term of course includes not only material differences, 

 but also those whose only demonstrable expression is 

 psychical. Thus, an increase in maternal affection is as 

 important and real a variation as the sharpening of a canine 

 tooth. 



It may also be useful to distinguish variations in size, 

 symmetry, number of appendages, and so on, from more 

 qualitative variations in chemical composition, such as 

 the appearance of a new pigment, but this distinction 

 is only a matter of convenience, as it is only a matter 

 of degree. 



Again, variations occur which may be called continuous ', 

 being merely minute increments or diminutions of certain 

 parental or specific characters. These are related to one 

 another much in the same way as are the successive stages 

 in the continuous growth of an individual. 



But other variations occur which deserve to be called 

 discontinuous. For, without the appearance of transitional 

 stages, marked variations crop up, reaching with apparent 

 suddenness to what must be called new and may withal 

 exhibit a measure of perfectness. 



That both kinds of variations occur is a fact of life ; the 

 possibility of both is probably a primary quality of 

 organisms ; but we are only beginning to know the relative 

 frequency of the two kinds and their respective limits, and 

 we know almost nothing as to their causes (see Bateson's 

 " Materials for the Study of Variation, 1894 " ). 



PRIMARY OR ORIGINATIVE FACTORS. 



What causes variation? This is the fundamental ques- 

 tion, but it is the least answerable. 



It is, indeed, an axiom or a truism, that changes in any 

 animate system are evoked by changes in the larger system 

 of which the organism forms a part. In other words, the 

 stimulus to organic change must always be ultimately 

 traceable to the environment, but this is implied in our 

 conception of living matter, and does not help us to under- 

 stand the immediate conditions which lead to change. 



In the absence of sufficiently precise data, we can do little 

 more than point out various possibilities : 



