188 Mr. H. Pender on the Magnetic 



potential of the fixed sector was from 100 to 130 C.G.S. elec- 

 trostatic units or 30,000 to 39,000 volts, while its distance 

 from the disk was only "5 cm. According to J. J. Thomson 

 (Recent Researches), the sparking-distance between planes 

 at a difference of potential of 100 C.G.S. electrostatic units 

 is *95 cm. Ebonite is not a perfect insulator, especially 

 for such high potentials. Hence the charge from the fixed 

 sector may have brushed across to the surface of the ebonite 

 facing the disk. This charge would be of opposite sign to 

 that induced on the metallic sectors, and the magnetic effect 

 of the rotating disk would therefore be approximately null. 



Present Investigation — Method and Apparatus. 



The method which has been employed in the present re- 

 search to put in evidence the existence of the magnetic action 

 of electrical convection is identical in principle with that used 

 by Cremieu in bis first experiments, viz., the measurement of 

 the current induced in a coil, on reversing the convection- 

 current due to a rotating charged disk. 



All experimenters who have worked on this problem have 

 been met by three difficulties, which have required no little 

 patience and considerable experimenting to overcome suc- 

 cessfully. These difficulties are : (1) the mechanical jarring 

 caused by the rapidly rotating disks ; (2) direct electrostatic 

 effects due to the charged disk ; and (3) external magnetic 

 disturbances. The present instance has been no exception to 

 this rule. In the first place, the galvanometer which it was 

 found necessary to employ was so sensitive, that it was 

 impossible to make any observations whatever except between 

 1.30 and 5 a.m., when the electric cars in the vicinity of the 

 laboratory had ceased making their regular trips. 



It is needless to describe the succession of trials which 

 were made to eliminate first one disturbing factor and then 

 another, and the various modifications made from time to 

 time in different parts of the apparatus. I shall describe the 

 apparatus only as it was used in the final quantitative 

 measurements. 



The Disk Apparatus. — This part of the apparatus was con- 

 structed several years ago under Prof. Rowland's direction, 

 shortly after his experiment in 1889, but had not been 

 previously used. A drawing of it is given in fig. 1 (p. 190). 

 Two micanite disks D D, \356 cm. thick and 30*4 cm. in 

 diameter, gilded on each side, and mounted on ebonite cores, 

 rotate about horizontal axes in the same line. On each side 

 of the disks are condensing-plates, formed by rings of tinfoil 



