DEVELOPMENTAL MECHANICS. 173 



few who have devoted themselves to the subject from more 

 productive activity. 



III. The Relations of Developmental Mechanics to the Other 

 Biological Disciplines. 



The branches of Biology hitherto recognized, viz., descriptive 

 zoology, anatomy -, embryology, and physiology, represent the essen- 

 tial prerequisites of developmental mechanics, for it is they 

 that teach us the facts in forms and processes, the causal 

 explanation of the latter being the province of the discipline we 

 are discussing. 



Because they depend on comparison of structure, anatomy and 

 embryology are also productive of causal information to the extent 

 that such comparison can take the place of experiment. 



This substitution cannot be a complete one for the logical 

 reasons presented above. Nevertheless, comparative anatomy 

 and comparative embryology are the means of ascertaining many 

 causal relations between the parts of organisms, and these rela- 

 tions, in so far as they rest on a sufficient mass of observations, 

 lack only the direct proof of artificial or natural experiment to 

 become certainty. In so far as these disciplines reveal causal 

 information, they are themselves developmental mechanics, and 

 inasmuch as they do and have done this to a very great extent, 

 they represent disciplines which are only historically separated 

 from developmental mechanics. 



The new character which these causal investigations have ac- 

 quired in recent times, and will continue to acquire, is the use 

 of analytical experiment, together with the endeavor to collect 

 together all causal information, and to raise causal investigation 

 to the dignity of a principal aim, an aim in itself. 



Thus phylogenetic and ontogenetic developmental mechanics 

 receive from the older branches of biology besides their prob- 

 lems much causal information, and still more guidance to 

 such information. The methods with which this knowledge has 

 been acquired will continue to be necessary to developmental me- 

 chanics even in future, since many causal problems are scarcely 

 accessible to experimental investigation, and since, moreover, 



