Chemical Composition of various kinds of Glass. 443 



with No. 5. Partly by friction with alcohol mixed with a little 

 hydrofluoric acid, partly by means of crocus-powder, I endea- 

 voured to remove the old surface, and as soon as possible after- 

 wards again tested the insulating-power of the glass. This 

 indeed showed some improvement, but so little that the glass 

 must still be considered utterly unserviceable for the purpose 

 of electrical insulation. 



I was accordingly driven to ask myself the question, whether 

 the rapid disappearance of electricity from certain kinds of glass 

 might not, in a far greater degree than is usually supposed, de- 

 pend on a passage of it through the substance of the glass itself. 

 This led to the following experiments. 



The specimens were subjected to the same test of their non- 

 conducting-power that has been above described, with the differ- 

 ence that now also the surface between the tinfoil coatings, pre- 

 viously left bare, was covered with shellac varnish. The relative 

 moisture of the air exhibited no material change. I obtained 

 the following times of insulation : — 





Specimen No. 1. . 



At least 12 hours. 



No. 2. . 



„ 1J hour. 



No. 3. . 



About 18 minutes 



No. 4. . 



„ 60 seconds. 



No. 5. . 



„ 40 „ 



It appears, then, that the insulating-power of the glasses had, 

 indeed, been increased by covering their surfaces entirely with 

 shellac ; but that increase appeared to bear a certain proportion to 

 the insulating-power displayed by the glasses while the surface 

 between the tin-foil coatings was still left bare. The increase is 

 therefore very inconsiderable for the inferior kinds of glass ; and, 

 indeed, the different sorts appear to differ in insulating-power as 

 widely as before. 



These experiments appear to me to justify the conclusion that 

 in the specimens of glass tested the electricity took its way princi- 

 pally through the substance of the glass itself, and that the nature 

 of the surface was of relatively subordinate importance. 



To verify more fully the correctness of this conclusion, I 

 made the following experiments with specimens 2 and 5, which 

 were plates of nearly the same thickness and circumference. In 

 the middle of the plate a circle of tinfoil, 7 centims. in diameter, 

 was fastened, and a similar one on the opposite side of the plate, 

 after which the whole remaining surface of the glass was var- 

 nished. The distance then between the tinfoil coatings was 

 the thickness of the plate (3 j millims.), reckoned through the 

 substance of the glass, but at least 12 centims. along its surface. 

 I now insulated the plate 5, put one (i. e. the upper) coating 



