ATTAINMENT OF VERY LOW TEMPERATURES. 7 



escaped from the tube as the gas expanded ; or the gas was com- 

 pressed directly by means of a pump into a glass or steel vessel, and 

 then part of it was allowed to escape directly through an orifice. In 

 the first case the gas exerted a pressure on the mercury, and on ex- 

 panding did work which was performed in giving the mercury a certain 

 velocity through the escape cock, and partly in overcoming fluid fric- 

 tion. In the second case, part of the gas which remained in the vessel 

 did work in driving the remainder through the orifice or cock, partly in 

 overcoming friction at the orifice, partly in giving to the gas an in- 

 creased velocity, which would quickly be dissipated in the formation of 

 eddy currents outside the orifice. 



If we were dealing with a perfect gas for which the simple law 



pv = const. 



were rigidly true, and if such a gas were allowed to expand, or were 

 compressed, adiabatically, the relation between pressure and volume 

 would be expressed by the equation 



p.^k __ const., 



where k is the ratio of the specific heat at constant pressure to the 

 specific heat at constant volume. If the gas were allowed to expand 

 adiabatically under the conditions mentioned above, where p^ and po 

 are the initial and final pressure, and T^ and To the initial and final 

 absolute temperatures. 





Since the heat capacity of the gas is always considerably less than 

 that of the vessel in which it is contained, the condition represented 

 above is probably never approached, though, as Olszewski's later ex- 

 periments show, a considerable degree of cooling can be effected. The 

 practical application of the principle is probably only possible on a 

 large scale. 



Olszewski's observation that when the initial pressure in the appa- 

 ratus exceeded a certain value the liquid always appears when the 

 pressure falls to the critical pressure, is not further discussed by him ; 

 the phenomenon can, however, be simply explained by means of an 

 Andrews diagram (fig. i). 



Suppose that the gas is compressed along the isothermal a a' a" a'", 

 and under the conditions determined by these points is, in successive 

 experiments, allowed to expand adiabatically. In each case liquid 

 should first appear at the point at which the adiabatic cuts the dotted 

 line enclosing the area which represents the conditions under which the 



