70 SCIENCE PROGRESS 



soluble crystalline salts, in mineral acids and in water, capable 

 of imparting to a mixture of chlorine and hydrogen the pro- 

 perty of temporary inertness towards light, was therefore very 

 probably ammonia. It was then discovered that the solutions 

 which previously had been found to induce the longest induc- 

 tion period were those which contained the greatest amount 

 of ammonia ; that ammonia-free water was incapable of ren- 

 dering chlorine inert ; moreover, that in relation to chlorine 

 and electrolytic gas, a dilute solution of ammonia corresponded 

 in every particular with the solution of crystallised barium 

 chloride previously examined. Now ammonia itself cannot 

 exist in the presence of chlorine, so that the actual substance 

 which induces the inertness of the mixture must be some 

 product of the interaction of chlorine and ammonia. 1 Nitrogen 

 chloride would appear to be indicated as the cause by the 

 following experiment. A dilute solution of ammonia was 

 saturated with chlorine and divided into two portions ; 5 cc. 

 of the first portion were added to an actinometer containing 

 chlorine and hydrogen, which was then exposed to light ; the 

 induction period was long. From the second portion the 

 chlorine was removed by exhaustion and 5 cc. of the purified 

 liquid was introduced into a similar actinometer; the induc- 

 tion period was very short. Hence it follows that the com- 

 pound which is the immediate cause of the induction period 

 belongs to the class of substances which are volatile and 

 readily removable by exhaustion on account of their slight 

 solubility in water. Moreover, it is destroyed by light and 

 heat. Nitrogen chloride fulfils these conditions. 



There was yet one outstanding fact which could not be 

 reconciled easily with the view that nitrogen chloride was the 

 sole cause of the induction period, the so-called phenomenon 

 of " deduction." When an actinometer containing electrolytic 

 gas and tap-water was exposed to light and shaken so as to 

 destroy the whole of the nitrogen chloride both in the gas and 

 in the liquid and the actinometer was left to stand in the dark 

 during several hours, it was generally found that the gas had 

 become inactive, i.e. that an initial inert period preceded steady 

 combination on re-exposure of the insolation vessel to light. 

 This behaviour could only be explained on the assumption 

 that the water contained some nitrogenous organic substance 

 1 Manchester Memoirs, Vol. XLIX. (1905), No. 13. 



