6 ART- 6— S. KUSAKABE : 



Figures in PI. IL show the twisting apparatus thus improved, 

 together with a rough sketch of the whole arrangement. Here it is 

 only necessary to remark that any couple whatever, positive or 

 negative, could be produced, without any increase of load to be sup- 

 ported by the knife-edge. 



The order of observation was as follows : — ■ 



1. To begin with, a specimen, whose dimensions were nearly 

 1x1x15 centimetres, was firmly clamped in horizontal position 

 and at right angles to the plane of the pulleys. 



2. The directions of the two mirrors attached to the specimen were 

 so adjusted that the two images of the vertical scale, whose 

 distance from the specimen was 2.716 metres, stood side by 

 side within the field of one and the same telescope, mounted on 

 a tripod. 



3. Equal weights, each ^ aI , were put on the pans Q t and Q 2 , 

 which, as a matter of course, gave no couple to the specimen. 



4. Zero-readings were taken firstly for the right image which was 

 reflected from the mirror attached close to the fixed end of the 

 specimen, then as fast as I could for the left image which cor- 

 responded to the free end. 



5. À definite number of pieces, which constituted the weight ^ M , 

 say \ M were taken oft' from one of the pans and put on the 

 other. The resulting couple was obviously equal to Mgr, if r 

 represent the radius of the twisting pully and g the value of 

 gravity. A time- record corresponding to this transposition of 

 weight was taken. 



G. After a certain definite time, the readings were taken for both 



images, as in the case of zero- readings. 

 7. Second transposition of weight ; the time recorded ; scale 



