Light on the Polarized Electrode. 429 



plates of coloured glass, one blue, the second yellow, and tlie 

 third red ; a strip of thick brown paper was pasted to the oppo- 

 site sides of each of these plates of glass, so as to form a nearly 

 cylindrical chamber cut by the plane of the glass. A cover was 

 placed over each of them ; and the chambers so formed could be 

 placed over the cell containing the platinum plates, the coloured 

 glass plates intervening between the sun and tlie platinum in the 

 outer cell. A great number of experiments were made with 

 these apparatus ; and in all the deviations of the galvanometer 

 were notably greater with the blue glass than with the yellow 

 or red, and, of the latter two, the yellow gave slightly greater 

 deflections than the red glass. 



This result is, I think, conclusive in favour of the eifect being 

 due to the chemical, not to the calorific rays of the sun, the more 

 so when we consider that the yellow allowed a far larger quantity 

 of light to pass than the blue glass. I may also add that I have 

 obtained a slight galvanometric deflection when diffused daylight 

 was allowed to impinge on the platinum plate, and when there 

 was no perceptible difference of temperature between the illumi- 

 nated and the non-illuminated plates. 



The superiority of the yellow over the red was not so strongly 

 marked ; and, considering that the yellow glass allowed much 

 more light to pass than the red, I am uot disposed to think that 

 there was any actual superiority in the former : the effects ob- 

 served with these two colours are, however, corroborative of the 

 effects not being due to the red or heating rays of the sun. 



I substituted for the water acidulated with sulphuric acid 

 (which may be regarded electrically as ])urc water with its con- 

 ducting power improved), muriatic and nitric acids ; the effects 

 were the same, but less marked with the nitric acid, probably 

 from its more completely depolarizing the plates. 



In a small number of experiments the following effect took 

 place. After a certain time of connexion with the galvanometer^ 

 the cover being over the apparatus, the sign of polarization 

 changed ; i. e. supposing the needle of the galvanometer to de- 

 viate to the left, and indicate that the exposed platinum was 

 positive, the needle would gradually return, pass the zero-point, 

 and be deviated to the right : when this was the case, on removing 

 the cover, the effect occasioned by the imjiact of light was a 

 return of the needle towards the zero-point, indicating an influ- 

 ence in the direction of the original polarization. On setting 

 aside the apparatus for twenty-four hours with the plates in me- 

 tallic connexion, and then repeating the experiment, the deviation 

 of the galvanometer was in the direction of the final polarization. 

 This apparent anomaly may have arisen from a conflict of two 

 classes of currents, the one arising from imperfect mixing or want 



