178 COMPARATIVE ELECTRO-PHYSIOLOGY 



Indeed it must be said that, however suggestive the 

 general theory of assimilation and dissimilation may have 

 been found, its abuse has often stood in the way of physio- 

 logical inquiry. The inquirer, when faced with any difficulty, 

 instead of attempting to surmount it by patient inquiry, was 

 tempted rather to evade it by invoking the aid of an hypothesis 

 which could be made with equal ease to explain a given 

 fact or its direct opposite. We must remember that in the 

 investigation of obscure problems, the danger is always, 

 instead of seeking an underlying law, to become satisfied 

 with the mere registration of phenomena, and by naming 

 these to imagine that they have been explained. The 

 resulting chaos in the present case has served to deepen 

 the impression that vital phenomena must always remain 

 capricious and mystical. 



But when we come to survey the facts that have been 

 described, we find the phenomena of response, however 

 diverse they may at first sight appear, to be in no way 

 governed by chance or caprice. They are, on the contrary, 

 definite and uniform under definite conditions. 



As regards the so-called current of rest in a naturally 

 isotropic tissue, of which one end has been subjected to 

 injury, we must remember that the effect of injury is one of 

 excitation, its sign, within limits, being contraction and 

 galvanometric negativity. But we have seen that when a 

 point is over stimulated, fatigue-changes appear which give 

 rise to a reversal of its normal sign of response, from con- 

 traction to expansion, from negative to positive (cf. fig. 64) 

 The change at death, in which contractile rigor passes into 

 post-mortem relaxation, is analogous to this. Thus when one 

 end of the specimen is merely injured, that end becomes 

 more or less persistently galvanometrically negative, the 

 current flowing away from it. But when the same end is 

 actually killed, the electrical change may be reversed, to one 

 of galvanometric positivity. 



In an isotropic tissue, then, we may, by moderate injury, 

 bring about a state of anisotropy, under which the uninjured 



