572 



SCIENCE. 



[N. S. Vol. XVI. No. 406. 



ting as high a vacuum as possible, and cold 

 is one of the best means of effecting the 

 desired exhaustion. All that is necessary 

 is to fill completely the space that has to be 

 exhausted with an easily condensable va- 

 por, and then to freeze it out in a receptacle 

 attached to the primary vessel that can be 

 sealed off. The advantage of this method 

 is that no air-pump is required, and that 

 theoretically there is no limit to the degree 

 of exhaustion that can be obtained. The 

 action is rapid, provided liquid air is the 

 cooling agent, and vapors like mercury, 

 water, or benzol are employed. It is obvi- 

 ous that when we have to deal with such an 

 exceptionally volatile liquid as hydrogen, 

 the vapor filling may be omitted because 

 air itself is now an easily condensable 

 vapor. In other words, liquid hydrogen, 

 collected in such vessels with the annular 

 space full of air, immediately solidifies the 

 air and thereby surrounds itself with a 

 high vacuum. In the same way, when it 

 shall be possible to collect a liquid boiling 

 on the absolute scale at about five degrees, 

 as compared with the twenty degrees of 

 hydrogen, then you might have the annu- 

 lar space filled with the latter gas to begin 

 with, and yet get directly a very high 

 vacuum, owing to the solidification of the 

 hydrogen. Many combinations of vacuum 

 vessels can be arranged, and the lower the 

 temperature at which we have to operate 

 the more useful they become. Vessels of 

 this kind are now in general use, and in 

 them liquid air has crossed the American 

 continent. Of the various forms, that va- 

 riety is of special importance which has a 

 spiral tube joining the bottom part of the 

 walls, so that any liquid gas may be drawn 

 off from the interior of such a vessel. In 

 the working of regenerative coils such a 

 device becomes all-important, and such 

 special vessels cannot be dispensed with 

 for the liquefaction of hydrogen. 



In the early experiments of Pictet and 



Cailletet, cooling was produced by the sud- 

 den expansion of the highly compressed 

 gas, preferably at a low temperature, the 

 former using a jet that lasted for some 

 time, the latter an instantaneous adiabatic 

 expansion in a strong glass tube. Neither 

 process was practicable as a mode of pro- 

 ducing liquid gases, but both gave valuable 

 indications of partial change into the liquid 

 state by the production of a temporary 

 mist. Linde, however, saw that the contin- 

 uous use of a jet of highly compressed gas, 

 combined with regenerative cooling, must 

 lead to liquefaction on account of what is 

 called the Kelvin- Joule effect; and he suc- 

 ceeded in making a machine, based on this 

 principle, capable of producing liquid air 

 for industrial purposes. These experi- 

 menters had proved that, owing to molecu- 

 lar attraction, compressed gases passing 

 through a porous plug or small aperture 

 were lowered in temperature by an amount 

 depending on the difference of pressure, 

 and inversely as the square of the absolute 

 temperature. This means that for a steady 

 difference of pressure the cooling is greats 

 er the lower the temperature. The only 

 gas that did not show cooling under such 

 conditions was hydrogen. Instead of being 

 cooled it became actually hotter. The rea- 

 son for this apparent anomaly in the Kel- 

 vin-Joule effect is that every gas has a 

 thermometric point of inversion above 

 which it is heated and below which it is 

 cooled. This inversion point, according to 

 van der Waals, is six and three-quarter 

 times the critical point. The efficiency of 

 the Linde process depends on working with 

 highly compressed gas well below the in- 

 version temperature, and in this respect 

 this point may be said to take the place of 

 the critical one, when in the ordinary way 

 direct liquefaction is being effected by the 

 use of specific liquid cooling agents. The 

 success of both processes depends upon 

 working within a certain temperature 



