310 Dr. J. Kerr on a new Relation between 



6. First experiment. — The pieces are arranged in the order 

 and manner just described (2, 3, 4). The spark -terminals are 

 fixed at a distance of 5 or 6 inches ; the polarizing Nicol is 

 laid with its principal section at 45° to the horizon ; and the 

 analyzer is turned, with the help of the neutralizing plate, 

 into the position of perfect extinction. No piece of the ap- 

 paratus is now touched, except the commutator, till the end of 

 the experiment. 



Looking through the polariscope, the observer closes the pri- 

 mary circuit. In about 2 seconds the light begins to reappear 

 through the dielectric, at its old place between the induction- 

 terminals, very faintly at first; but it brightens continuously 

 for 10 seconds, 20, even 30, till it is almost brilliant. When 

 the primary circuit is now broken by the commutator, the 

 light fades away continuously, at first rapidly, then more 

 slowly, to perfect extinction. The time that elapses (in the 

 latter part of the experiment) between the opening of the pri- 

 mary circuit and perfect extinction in the polariscope depends 

 very noticeably upon the intensity and duration of the electric 

 action, increasing as these increase. 



7. The light thus restored by electric action cannot be extin- 

 guished again, at any stage of the experiment, by any rotation of 

 the analyzer either way. 



8. Second experiment. — The polarizing Nicol is laid with its 

 principal section either horizontal or vertical ; the analyzer is 

 turned into the corresponding position of perfect extinction ; all 

 the other arrangements and the procedure are as in the first ex- 

 periment. There is now no regular effect obtained in the polari- 

 scope. In many cases, indeed, even when the strongest electric 

 action (5) has been kept up for 20 or 30 seconds, any recovery 

 of the light is very doubtful, rather a guess than a perception. 



Small effects do sometimes present themselves ; but they are 

 trifling and irregular in comparison with those obtained in the 

 first experiment. They are probably due to known causes, such 

 as imperfection of adjustments, irregularities of molecular struc- 

 ture in the dielectric, possibly also slight changes of tempera- 

 ture. If a small and irregular allowance be made for one or 

 more such disturbing influences, there is now no effect in the 

 polariscope. 



9. Third experiment. — Distance of the spark-terminals small 

 (say, 2 inches), the arrangements otherwise as in the first expe- 

 riment. The electric action is kept up for a minute or more, till 

 the intensity of the restored light is certainly constant. The 

 spark-terminals are then separated all at once to a distance of 6 

 inches, and in a second or two there is an evident increase of 

 effect in the polariscope. Simple arrangements will be described 



