380 PHYSIOLOGICAL DEVELOPMENT. 



wise differentiated, it must equally happen that as fast 

 they become channels along which molecular disturbances 

 travel, the parts they connect become physiologically inte- 

 grated, in so far that a change in one initiates a change in 

 the other. We may dimly perceive that if portions of what 

 was originally a uniform mass having a common function, 

 undertake subdivisions of the function, the molecular 

 changes going on in them will be in some way complemen- 

 tary to one another : that peculiar form of molecular motion 

 which the one has lost in becoming specialized, the other has 

 gained in becoming specialized. And if the molecular motion 

 that was common to the two portions while they were undiffer- 

 entiated, becomes divided into two complementary kinds of 

 molecular motion; then between these portions there will be 

 a contrast of molecular motions such that whatever is plus 

 in the one will be minus in the other ; and hence there will be 

 a special tendency towards a restoration of the molecular equi- 

 librium between the two: the molecular motion continually 

 propagated away from either will have its line of least resist- 

 ance in the direction of the other. If, as argued 

 in the last chapter, repeated restorations of molecular equili- 

 brium, always following the line of least resistance, tend ever 

 to make it a line of diminished resistance; then, in propor- 

 tion as any parts become more physiologically integrated by 

 the establishment of this channel for the easy transmission 

 of molecular motion between them, they may become more 

 physiologically differentiated. The contrast between their 

 molecular motions leads to the line of discharge; the line of 

 discharge, once formed, permits a greater contrast of their 

 molecular motions to arise; thereupon the quantities of 

 molecular motion transferred to restore equilibrium, being 

 increased, the channel of transfer is made more permeable; 

 and its further permeability, so caused, renders possible a still 

 more marked unlikeness of action between the parts. Thus 

 the differentiation and the integration progress hand in hand 

 as before. How the same principle holds through- 



