2 Vol. XXXVII.. Art. 1.— K. Fuji : 



galvanometer, whereas Cremer employed his string electrometer. 

 Although the metliod of experiments used by Gotch was certainly 

 the most ingenious at that time, the curves obtained are not 

 sufficiently accurate to be aiial,ysed quantitatively. The method of 

 Schönlein is inferior to tliat of Gotch, for the form of the discharge 

 curves obtained l»y liis metliod cannot fail to be mucli affected by 

 the fatigue of the tissues. As the capillary electrometer, the string 

 galvanometer and the string electrometer cannot faithfully follow 

 so quick a change in the electromotive force as happens in the case 

 of the discharge plienomena in question, tlie curves olttained by 

 til em are only qualitative. 



In April 190G, in conjunction with ]Mr. S. Ginuma, now 

 I(jahuhahusJù, the autlior began his experiments on the same 

 subject by means of an oscillograpli in the Physiological Institute 

 of the Imperial University of Tokyo. In the course of this investi- 

 gation, his collaborator went to Europe, and the author was then 

 obliged to continue the experiments alone. 



The electric hsli used Ijy us was one of the electric rays called 

 Astrapa japonica, which was brought from Misaki, a fishing town 

 in Sagami peninsula. Perhaps on account of the imperfectness 

 of our aquarium, the fish could not be kept in a healthy state ex- 

 cept for a few months in Spring and Autumn; and in these most 

 favourable periods, even a single fish sometimes could not be ob- 

 tained. ()n account of these and other hindrances, and although 

 lengthy and exhaustive experiments were continued up to 1910, 

 the problems first proposed could not l)e thoroughly investigated. 

 The results so far obtained l)y these experiments arc now presented 

 in this paper. 



With regard to the form of the discharge curve which will be 

 discussed in § IV., it is very interesting to note that some physio- 

 logical phenomena can be explained from the standpoint of the 

 theory of probabihty. It is a known fact that many quantities 

 characterising a class of animals or plants are subject to individual 

 variations; and if the observations are taken on a sufiiciently large 

 number of individuals of the same class, the values representing any 

 of these characteristics distribute themselves, among the number of 



