REQUIRED TO PRODUCE DISCHARGE THROUGH VARIOUS GASES. 



887 



closed. On allowing the mercury to rise in the pump reservoir, the gas in the 

 bulb 6 was compressed, and forced the surface of the alloy down to f. The gas then 

 bubbled up through the bulb a, and was compressed into the sparking vessel. When 

 the mercury reservoir was lowered again, the gas bubbled back again through the 

 alloy, now in the bulb b. 



Fig. 4. 



V 



This process was repeated as often as required. 



It was not possible to pass all the gas through the alloy at each operation ; yet by 

 giving time for the diffusion of the gas compressed into the sparking vessel, so as to 

 allow it to become of uniform composition throughout, before being again passed 

 through the absorbent, a very efficient purification could be effected by a few 

 passages of the gas through the alloy. When as much had been drawn into the pump 

 reservoir as circumstances allowed, the amount remaining behind in the sparking 

 vessel was not more than one-fifth of the whole. Thus, supposing all the oxygen to 

 be taken out of that part of the gas which had actually bubbled through the alloy, 

 only | would remain after the first operation, -$ after the second, yjj after the third, 



so on. 



When it was desired to measure the spark potential in the purified gas, the tap c 

 opened, so as to make the pressure the same in the manometer as in the sparking 



Vr.xsi'l. !(' this ll.nl Hot !".!! lion.', tin- plv^mr illlf t<> t 1 1- < '"! U 1 1 1 1 1 ")' Ji 1 1' -V \\ ' 'II l< 1 



3 D 2 



