CONDITIONS OF CHEMICAL CHANGE 69 



distilled water had been used to dissolve the hydrogen chloride 

 produced in the interaction. An aqueous solution of a par- 

 ticular specimen of crystallised barium chloride was found to 

 be exceptionally effective in rendering chlorine inactive. Our 

 attention was next turned towards the discovery of all the 

 possible methods by which such aqueous solutions could be 

 rendered incapable of imparting inertness to electrolytic gas. 

 It was found that this could be best accomplished by the simple 

 expedient of saturating the aqueous solution with chlorine and 

 then boiling the solution until as much as possible of the 

 chlorine had escaped. Attempts were next made to remove 

 the residue of chlorine from a solution which had been treated 

 in the above manner without at the same time introducing the 

 obscure agent which is capable of rendering electrolytic gas 

 temporarily insensitive to light, in order to provide ourselves 

 with the means of testing conclusively whether the agent in 

 question be a transitory and communicable property of the 

 salt — some substance produced from the pure salt — or simply 

 some foreign impurity contained in the original sample of the 

 salt. One of the methods adopted for removing the last trace 

 of chlorine from the boiled solution was to add a drop or two 

 of a solution of potassium iodide and then to remove the liberated 

 iodine with a solution of sodium thiosulphate ; the solution 

 from which the chlorine had been thus entirely removed was 

 incapable of imparting inertness to electrolytic gas nor did it 

 acquire the property on standing or on heating. We were 

 thus driven to the conclusion that the inhibitive agent was 

 neither an acquired property of the salt nor a substance 

 developed from the pure salt but some unknown, widely dis- 

 tributed impurity contained in the original sample of the salt. 

 Just as we were on the point of starting a series of experi- 

 ments with the object of concentrating and isolating the 

 impurity, a pure accident disclosed its identity. It was thought 

 that a method better than that just described of removing the 

 residue of chlorine from the boiled solution which had been 

 saturated with chlorine would be to add just sufficient ammonia 

 to destroy the chlorine ; on introducing the liquid treated in this 

 way into an actinometer which already contained a mixture 

 of chlorine and hydrogen and then exposing the actinometer 

 to light, an inert period of abnormally long duration was 

 observed. The widely distributed impurity present in many 



