DEVELOPMENTAL MECHANICS. 163 



single component which we intended to change is affected, are 

 we in a position to draw a definite causal conclusion from a 

 single experiment. 



This conviction or insight will only rarely be obtained from 

 experiments on organisms. Hence it so often happens that 

 when we believe we have experimented under the very same 

 conditions and in the same manner as on a former occasion, 

 we nevertheless obtain different results. So long as we do not 

 arrive at the same result, at least after several repetitions of 

 the same experiment, we must not permit ourselves to draw 

 any conclusion whatsoever. And now that we are in the first 

 stages of our investigations, without having any survey of the 

 modi operandi which may occur, it will often be necessary to 

 use as many methods as possible in experimenting on the same 

 subject ; and only when these different experiments point to 

 the same causal connection should we assume that this is the 

 true one. 



With the aid of such experiments we are in a position on 

 the one hand to test the relationships which are determined by 

 comparative study of the normal forms, and on the other hand 

 to obtain yes, to extort an answer to newly arising ques- 

 tions. 



Before we can establish the causal modi operandi according 

 to their qualities, we must first determine the parts between 

 which formative operations take place, i.e., we must determine 

 the " locality" of the formative operations. With reference to 

 the single circumscribed structure or part, this means that we 

 must ascertain whether the causes of its formation lie within 

 itself or whether external influences are necessary to its for- 

 mation. 



The role which the different causes that take part in a for- 

 mation play in its production may be a very unequal one. 



Inasmuch as some singular notions and terminology have gone 

 abroad concerning causes of different dignity, it seems proper 

 in this place to go somewhat into details for the sake of paving 

 the way towards greater uniformity of opinion. 



All the components whose temporary and local coincidence is 

 necessary to produce a certain effect, constitute in their 



