MEMOIR XX.] THE ALLOTROPISM OF CHLORINE. 287 



fluenced by the sun has obtained the quality of effecting 

 this decomposition subsequently, to a measured extent, 

 even in the dark. Not to anticipate what I have to 

 offer on this point, I shall now proceed in the first place 

 to establish the various facts connected with the decom- 

 position in question. 



Having provided a number of small glass vessels, con- 

 sisting of a bulb and neck of the capacity of from 1.5 to 

 2 cubic inches, I filled them with a solution of chlorine 

 in recently boiled w^ater, and inverted them in small 

 glass bottles containing the same solution, as shown ^_^ 

 in Fig. 44. With these bulbs the following exper- (~ \ 

 inients were made : 



I. An aqueous solution of chlorine does not de- 

 compose in the dark. 



One of the bulbs was shut up in a dark closet, 



and kept there for a week, being examined from Fi s- 44 - 

 time to time. No decomposition was perceptible, for no 

 gas collected in the upper part of the bulb. 



II. An aqueous solution of chlorine decomposes in the 

 light. 



One of the bulbs was placed in a beam of the sun re- 

 flected into the room by a heliostat. For sixteen min- 

 utes no change was perceptible, then small bubbles of 

 gas made their appearance; they increased in quantity 

 for a time, but finally the speed of decomposition became 

 uniform. On analysis by explosion with hydrogen, after 

 washing out any chlorine contained in it, this gas was 

 found to contain 97 per cent, of oxygen. 



III. The rapidity of this decomposition depends on 

 the quantity of the rays and on the temperature. 



In various repetitions of these experiments on differ- 

 ent days I soon convinced myself that the rate of evo- 

 lution of the oxygen depended on the quantity of the 

 rays. Among other proofs I may mention this: After 



