92 Dr. J. Kerr's Electro-optic 



(8). It would be rash to assort that there is no considerable 

 discharge of electricity through the coll in the course of that 

 experiment. On the contrary, I think that when the wires 

 are in position and the machine working, there is a certain 

 quantity of convectivc discharge through the liquid. But the 

 present observation shows, to a certainty, that the shell of bi- 

 sulphide in the cell, though not more than one eighth of an 

 inch thick, is able to keep the two inner balls at a large 

 difference of potentials, a larger difference than would be 

 supported under similar conditions by more than an inch of 

 air. The liquid is therefore a good insulator ; and the resto- 

 ration of the light by electrization is due, in all probability, to 

 electrostatic inductive action through the liquid. Up to this 

 point I have merely given a revision of observations made long 

 ago with the old plate cell, and published in the second of my 

 former papers. 



11. Extinction-bands in CS 3 . — The cell is charged with clean 

 carbon disulphide, and the pieces all stand as in the diagram 

 of (6), the principal section of the first Nieol being inclined 

 at 45° to the horizon, and the second Mcol being at pure ex- 

 tinction. The correctness of the arrangements is tested by a 

 repetition of the electro-optic experiment with the hand com- 

 pensator (9). Matters being thus arranged, I begin now by 

 attaching a weight of some pounds, say eight or nine, to one 

 of the fixed compensating slips B (6). The experiment begins 

 thus with a strong permanent restoration from extinction in 

 the polariscope, the restoration being due to vertical tension 

 of glass, which is here optically equivalent to horizontal com- 

 pression of glass, and therefore optically contrary to the 

 electric action (9). It should be remembered that the object 

 now restored in the polariscope is a fine vertical streak of 

 flame-light, passing through the centre of the electric field, 

 not encroaching on either ball, but projecting well above the 

 balls, and also well below. Things being thus prepared, the 

 observer sits at the polariscope, and the machine is worked at 

 a moderate rate. 



The first thing observed is a broad horizontal band, very dim 

 and ill-defined, which crosses the flame in the axal part of the 

 electric field. As the potential of prime conductor and induc- 

 tric ball rises, the band comes out more and more definitely, 

 and darkens by degrees till it is perfectly black, every trace of 

 that part of the flame having disappeared. As the potential 

 still rises the flame begins to reappear in a faint speck or patch 

 at the centre of the band. The patch brightens and widens 

 gradually till the one band along the axis is clearly broken up 

 into two, lying symmetrically on opposite sides of the axis, 



