Hydrogen and the Spectrum of Aqueous Vapour. 341 



seals for these joints, having satisfied myself that all other 

 forms of connexions consume time and are worthless. The 

 tubes were mounted on a board, and after the joints had been 

 adjusted, the final adjustment to the gas-holder could be 

 readily and safely accomplished by means of a rack and 

 pinion which raised or lowered by slow degree the gas-holder 

 G, fig. 2. This drying arrangement could be exhausted to a 

 high degree, and would hold the rarefied gas as long as 

 desired. The method of filling the tubes was as follows : — 

 The spectrum-tube was exhausted to about 1 millim., having 



Fi°-. 2 



been repeatedly filled with hydrogen. This operation was 

 continued until tlie hydrogen spectrum appeared. Finally, 

 the tube was exhausted to the x-raj stage by long heating 

 and by employing a condenser-discharge. Then the dried 

 gas was admitted until a pressure of about *1 millim. remained 

 in the tube. The vessels with palladium tubes were carefully 

 exhausted to the .x-ray stage (and beyond, a spark preferring 

 to jump six inches in air to passing through the tube). When 

 these .spectrum-tubes, with the palladium adjunct, were con- 



