SILENT DISCHAKGE OF ELECTRICITY ON OXYGEN AND OTHER GASES. 215 



per cent, of oxygen ; that is to say, 9997 and 99-96 per cent, of the gaa 

 was absorbed by melted phosphorus in experiments made upon the two 

 samples. 



When the apparatus is not in use the taps t\ t^, t^ are closed, and 

 the traps C and D filled with mercury to prevent the entrance of air. 



Oxygen may be delivered from A into any vessel by connecting it to 

 P, exhausting it, t^ being closed and Q being clamped to prevent the 

 mercury from rising and filling C, and subsequently opening t^, when the 

 gas will flow from A into the exhausted vessel. If f be well ground it 

 will resist the passage of air sufficiently to permit this to be done. 



If a delivery tube be attached to P, all air may be driven from it by 

 flushing it with mercury before proceeding to deliver the oxygen in the 

 usual mancer. Thus waste of pure gas is avoided. 



The supplementary tap f and mercury trap D are provided in order 

 that accidental breakage of E on the application of heat (when fresh 



VA-;; 



supplies of oxygen are about to be introduced into A) shall not admit air 

 to the stock of oxygen already in A. When A, partly empty, is to be 

 replenished, t^ is closed, t'^ opened, and heat is applied to E, D being filled 

 with mercury to the level S, through which the gas is permitted to 

 escape. When E is thoroughly heated and a steady evolution of gas 

 has set in, the oxygen is delivered into A. Thus if the replenishment of 

 A be not too long delayed, no loss of time results from accidents to E, 

 which can at any time be replaced and exhausted, whilst the oxygen 

 remaining in A is still available for use, if the taps have been properly 

 ground and are thoroughly lubricated. Of course the trap D, like C, 

 must be closed by filling it with mercury at all times when escape of 

 gas from E is not desired. 



When potassium chlorate is used as the source of oxygen, breakages, E, 

 are frequent. Silver oxide is much better, but it is troublesome to obtain 

 it perfectly free from carbon dioxide. This has led us to employ a 

 mixture of the chlorates of sodium' and potassium in molecular propor- 



' Chlorate of sodium ia apt not to be pure ; it should be carefully examined before 

 it is used. 



