60 ART. 4. — K. HOXDA AIS^D Ï. TERADA. 



of which fitted to agate cups on the arms of a F-shaped bar (a). 

 This bar could be adjusted to any desired position, such that the 

 axis of the mirror would be pressed by the plane side of the 

 carriage perpendicularly on a point of the circumference of the 

 torsion wheel (t) rigidly fixed to the specimen. To adjust the 

 pressure properly, a sliding weight {b) was put on a pin (p) pro- 

 truding from the carriage. 



Two vertical scales oO cm. long were erected at a horizontal 

 distance of 6.45 m. in front of the mirror in the same vertical 

 line; one at the same height as the mirror, and the other about 

 one meter above it. The image of the scales was observed with 

 a telescope at the same height as the mirror. An initial couple 

 was applied and the mirror was so adjusted that the image of 

 the lower scale was observable with the telescope. A suitable 

 weight was then chosen for giving an additional couple, which 

 twisted the rod to such an extent that by the consequent rotation 

 of the mirror, the image of the upper scale just appeared in the 

 field of the telescope. Tliis weight was 500 gr. The readings of 

 the two scales corresponding to the removal and to the addition 

 of the additional weight, were taken. When properly adjusted, 

 their difference was fairly constant for repeated observations ; the 

 mean of several observations was always taken as the deflection 

 due to the weight. Next, a field was applied and the procedure 

 repeated ; in this way, we obtained the values of deflections cor- 

 responding to the difierent fields, the demagnetization being, of 

 course, carefully made before each experiment. The difference of 

 the deflections in a certain field and in no field gives the change 

 of rigidity. 



The sensitiveness of the present arrangement was such that 

 a deflection of 1mm. of the scale reading corresponded to a twist 



