a Dielectric in an Electrostatic Field. 199 



by former investigators, and from experiments of his own, 

 believes that these observed changes of volume and length 

 are to be accounted for by electric compression of the elec- 

 trodes and the heating effect of the electric charge. In 

 support of this opinion, that at least part of the effect was 

 due to the former, it should be noted that in every case there 

 has been an abrupt change of potential at the surface of the 

 dielectric. This would necessitate the action of free charges 

 on the surface of the glass or liquid, and the attraction of 

 these charges would produce changes in volume. In general, 

 from the arrangement of the apparatus, it would cause an 

 apparent expansion. 



My principal aim in these experiments was to prevent any 

 free or apparent charges on the surface of the dielectric and 

 so obviate this criticism. As for temperature changes, they 

 were said to be slower in action than the electrical strains, 

 and so separable. If a solid dielectric, as glass, were charged 

 by electrodes not in contact with it, and if the spaces inter- 

 vening were filled with a liquid non-conductor of the same 

 or nearly the same specific inductive capacity, it is evident 

 that the potential gradient between the charged plates would 

 be uniform and there could be nothing like a free or apparent 

 charge on the surface of the glass. After a number of 

 apparatus were tried, the following was found best to answer 

 the purpose. 



A brass tube (fig. 1, p. 200) 61 centim. long, 4centim. inner 

 diameter, and 4 millim. thick, was screwed and then soldered 

 into a heavy iron plate. This, in turn, was fastened by 

 screws into a ring of hard wood. Around the base of the 

 tube a collar of brass 3 - 5 centim. high was soldered. The 

 glass tube to be tested was then passed over the inner cylinder 

 and sealed with sealing-wax to a height of 4 centim. This 

 tube had a length of 57*5 centim., inner diameter 5 - 2 centim., 

 mean thickness 2*5 millim. A third tube was fitted over these 

 two. It consisted of three pieces, the middle part of brass, 

 44 centim. long, 3'5 millim. thick, 6*3 centim. inner diameter, 

 and two end-pieces of glass which fitted tightly over the 

 brass tube as shown in the figure. The three were then sealed 

 together, and the whole, in a brass ring screwed to the iron 

 base. This arrangement of glass and metal was for the 

 purpose of insulating the brass part when the latter was 

 charged. To prevent leakage the outer surface was painted 

 with shellac. Great care was taken, when putting the apparatus 

 together, to make the axes of the three cylinders coincide. 

 The spaces between the cylinders could be rilled with various 

 liquids as desired. A thin layer of mercury, poured between 



P2 



