THE LIQUEFACTION OF GASES 51 



Liquid air can be obtained in any quantity 

 by the expenditure of power, and the necessary 

 apparatus has become part of the usual equip- 

 ment of physical and chemical laboratories. By 

 this means regions of temperature before quite 

 inaccessible have been opened up to investiga- 

 tion, and the use of liquid air promises to be of 

 increasing advantage in many departments of 

 research. It would, of course, be possible to 

 drive an engine by means of liquid air, but such 

 a process would be very uneconomical. The state- 

 ments, which have sometimes appeared in the 

 daily papers, announcing impending revolutions 

 in methods of obtaining cheap power by the 

 application of liquid air, have originated from an 

 imperfect comprehension of the problems involved. 



When air had been successfully liquefied, 

 hydrogen was obviously the next gas to be 

 attacked. Thomson and Joule's porous plug 

 experiments had shown that, at ordinary 

 temperatures, hydrogen suffers a heating effect 

 on free expansion. It was therefore useless to 

 attempt to liquefy it by regenerative cooling 

 alone. But, just as the cooling effect in the case 

 of air increases as the air is subjected to a pre- 

 liminary cooling, so in hydrogen, if it be first 

 cooled, the Thomson-Joule heating effect first 

 diminishes and then is reversed, becoming a 

 cooling effect. This reversal was shown by 

 Olszewski to take place about 80° below the 

 Centigrade zero. Dewar then subjected hydrogen 

 to a preliminary cooling in liquid air boiling in a 

 vacuum at a temperature of — 205°, and afterwards 

 forced the hydrogen through a regenerative coil 

 under a pressure of 180 atmospheres. 



