184 Mr. EL Pender on the Magnetic 



the magnetic circuit through ihe coil. The whole distance 

 from plate to plate \Aas 8 mm., the disk revolving in the 

 centre of this opening. A brush rubbing on a brass ring 

 mounted on the insulating hub and connected to the metal 

 disk, served to connect the disk with one terminal of a high 

 potential battery (5000 volts), the other terminal of which 

 was earthed. On charging or discharging the disk, the 

 condensing-plates being earthed, on the assumption that a 

 moving charge produces a magnetic action, a current should 

 be induced in the coil, and the needle of the galvanometer 

 deflected. The calculated effect of a single charging or dis- 

 charging was too small to be detected, so that a combined 

 interrupter and commutator was employed. This instrument 

 served to charge and discharge the disk a number of times a 

 second, and at the same time to commutate any alternating 

 current which might have been induced in the coil. The 

 galvanometer should therefore have shown a permanent 

 deflexion when the commutator was kept running at a con- 

 stant speed. No such deflexion was observed, though the 

 calculated deflexion was from 20 to 50 mm. 



Four series of experiments were performed. In the first 

 was used a solid aluminium disk 0*5 mm. thick. This was 

 charged to 5000 volts, but it immediately discharged itself by 

 sparks across the small gap between the disk and condensing- 

 plates. To prevent this, the latter were covered with glass- 

 plates, lacquered, 2 mm. thick. This, however, left a space 

 of only 1'75 mm. between the disk and glass surface. It 

 was suggested by Pellat that perhaps the charge brushed 

 across from the disk to the glass surface, thus leaving the 

 disk uncharged. This was tested by Cremieu for the disk at 

 rest, and found not to be the case. But it is impossible to 

 make a disk of aluminium only 0*5 mm. thick and 37 cm. in 

 diameter run perfectly true at a speed of 100 revolutions per 

 second. Consequently, when the disk teas rotating the interval 

 between it and the glass surface must at certain points have 

 been considerably less than 1*75 mm., and therefore the 

 above objection still holds. This is the explanation of the 

 negative results of Cremieu's experiments which was given 

 by Professor Rowland. 



In his second series of experiments Cremieu used an alu- 

 minium disk 1 mm. thick covered w T ith a thin coating of 

 caoutchouc, the condensing-plates being left bare. Again 

 negative results were obtained. When it is remembered that 

 the distance between the disk and the condensing-plates was 

 only 3'5 mm., and that an aluminium disk of the diameter 

 used, even though 1 mm. thick, will not run perfectly true, it 



