OF ARTS AND SCIENCES. 



883 



This method of observation, however, was far from satisfactory ; the 

 passage of the sparks was very capricious, and it often seemed as if the 

 armature was of no advantage ; although, if the trials were extended 

 over a sufficiently large interval, a gain was shown in using the arma- 

 ture. I speedily resolved to make use of one of Sir William Thomson's 

 new quadrant electrometers, and to measure the difference of potential 

 of the terminals of the secondary coil directly. The coils were arranged 

 as previously, and by means of a peculiarly constructed key the elec- 

 trometer terminals were connected with those of the secondary coil. 

 A small grove cell of the Elliott pattern was used. This cell, how- 

 ever, was very much weakened by a shunt. The following tables show 

 the results which were obtained ; the deflections are expressed in the 

 divisions of the electrometer scale. 



WITHOUT THE ARMATURE. 



