Electricity and Light. 345 



there is a good virtual compression of the dielectric at the centre 

 of the held, there is an effect of the same kind, a virtual com- 

 pression in the same direction, beginning to manifest itself all 

 along the equator, but more faintly as the distance from the 

 centre of the field increases. Out of this equatorial line, and 

 well back from the wires, the direction of apparent compression 

 of the dielectric changes from point to point of the field, and is 

 at some points vertical, perpendicular to the line joining the 

 terminals. Only at points very close to the induction-wires does 

 the compensator fail decisively to extinguish or greatly weaken 

 the restored light. 



20. In connexion with the preceding facts, I may mention 

 that under certain conditions, in the fifth experiment, the com- 

 pensator does not extinguish the restored light all at the same 

 time, but produces a very dark broad band, which descends from 

 the outer parts of the field towards the centre as the tension of 

 the compensator increases. This happens, for instance, though 

 not very regularly, when the electric force is very intense from 

 the beginning and the optical effect is just beginning to show 

 itself; it happens also when the lines of force are inclined to the 

 horizon, so that the fine vertical band of flame seen through the 

 centre of the dielectric crosses the equator of the field at about 

 45°. In these cases there appears to be an unusually rapid 

 variation of birefringent action through the part examined of 

 the electric field. 



I have now done with the dielectric of plate glass. Half a 

 dozen other solids were tried ; and I will conclude with a short 

 account of the only two of them which gave results worth men- 

 tioning. The great difficulty was to get a sufficiently strong- 

 superficial insulation, the masses being too small. In only 

 one case, that of Iceland spar, was the dielectric perforated by 

 discharge without giving a clear effect; but I think the crystal 

 had received a predisposing flaw in the operation of boring. 



21. Dielectric of Resin. — This is a piece very similar to the 

 dielectric of plate glass in form, size, and adjuncts. A quantity 

 of clear amber-resin is kept at a gentle fusion for some time till 

 it is free from air-bubbles ; it is then poured into a suit- 

 able mould and left to cool. Two thick stocking-wires of steel, 

 previously fixed in the mould, along one line through its ends, 

 remain imbedded as induction-terminals in the solid resin, leaving 

 less than ^ of an inch of clear dielectric between them at the 

 centre of the block. The polarized light enters and leaves the 

 central part of the block through small squares of thin plate 

 glass, which are in optical contact with the resin and parallel 

 to each other. 



This dielectric, even the best specimen of it, is far inferior 



