AffARAfttS 8? 



of curves was obtained, in which it appears that air steadily 

 diminishes in volume as pressure is applied at common tempera- 

 tures, while carbon dioxide at temperatures considerably above 

 31 C. imitates it pretty closely. If the temperature, however, 

 approaches within a few degrees of 31 a sudden diminution of 

 volume is observed, which is indicated by a sudden change of 

 direction of the curve, becoming vertical at anything below 31 

 with a pressure about 75 atmospheres. 



There is therefore a critical point of temperature above 

 Which carbon dioxide gas cannot be reduced to liquid by pres- 

 sure. This has been shown to be true of all other gases, and the 

 reason why oxygen, hydrogen, and some others had not been 

 liquefied by pressure up to that time was that the temperature 

 emplpyed in each case had been above the critical point of the 

 gas. The critical temperature for oxygen is about -118 C., 

 while that of nitrogen is -146 C. Cooling the gas below these 

 temperatures is therefore an essential condition for their lique- 

 faction, and in the case of oxygen this result was attained, for 

 the first time at the end of 1877 by the French physicist Cailletet 

 and the Swiss professor Pictet, by two distinct methods. 



But for the continuous production of the liquid from the gas 

 yet another principle was necessary. Experiments made nearly 

 seventy years ago by Dr. Joule of Manchester, in conjunction 

 with William Thomson (afterwards Lord Kelvin), resulted in 

 the discovery that if a gas is caused to expand so as to do 

 external work cooling results. Joule and Thomson caused a 

 stream of compressed gas to pass through a long copper spiral, 

 immersed in water, at constant known temperature. The gas 

 then escaped through the pores of a plug of compressed cotton 

 wool, and its temperature was noted. In every case, except 

 hydrogen, a reduction of temperature was observed which, 

 though actually small in amount, was the greater the lower the 

 temperature of the gas before passing through the plug. Thus 

 the effect on carbon dioxide is shown in the following figures : 



Temperature Degrees Cent. 



Before escape through plug . . 12-8 19-1 91-5 

 Reduction of temperature . . 1-207 1-144 0-69 

 The amount of cooling is proportional to the difference of pressure 

 before and after release, and inversely as the absolute temperature. 

 At low temperatures it has been found that hydrogen behaves 

 like other gases. These results afford information as to the 

 internal constitution of gases, for, on allowing a gas to expand, 



