the Boiling Temperature of Hydrogen. 209 



those given now, and were based npon many careful expe- 

 riments. Considering that the temperatures extrapolated 

 must have been determined a little too low, for the reasons I 

 have stated above, the numbers previously communicated may 

 even be nearer the true data. 



I showed above how the expansion method enables us to 

 determine the dependence of the temperature upon the pressure 

 of the gas liquefied, even though the cooling agents do not 

 allow us to reach the critical temperature of the gas experi- 

 mented with. It would be possible to make the following 

 objections. First, we might ask how it is possible to know 

 whether the gas, expanded from a temperature higher than 

 the critical temperature, to a pressure of 20 atmospheres or 

 of 1 atm., really assumes a temperature equal to the tempera- 

 ture of the liquefied gas, under one or other of these pressures. 

 And then, if this were granted, whether the platinum thermo- 

 meter described is rapid enough in its indications to assume 

 and to mark the temperature of the surrounding gas at the 

 moment of the expansion. 



To ascertain whether these objections had any real founda- 

 tion, I have done just what I did for the determination of the 

 critical pressure of hydrogen, viz., I performed a series of 

 analogous experiments with oxygen, whose critical and boiling 

 points, and vapour -pressures at several different pressures I 

 have determined in a former paper * , using the hydrogen 

 thermometer. The experiments were performed in a similar 

 manner, with this difference, that for the iron flask containing 

 hydrogen under 170 atm. I substituted another flask, contain- 

 ing oxygen under 110 atm.; to coo] the steel vessel containing 

 the platinum thermometer I used liquid ethylene, boiling 

 under atmospheric pressure ; then I cooled the oxygen, which 

 was under a pressure of 100 to 110 atm., to a temperature 

 which was about 16° higher than its critical temperature, and 

 expanded it to 50*8 atm. (critical pressure), to 32*6 atm., to 

 19 atm., to 10*2 atm., and to 1 atm. The measurement of the 

 resistances of the platinum thermometer was done more quickly 

 in these experiments than in those with hydrogen ; for, 

 knowing from my former experiments the relation between 

 the temperature and the pressure of liquefied oxygen, I was 

 enabled to determine beforehand with very near approxima- 

 tion the resistances which ought to be inserted in the 

 resistance-box. 



* Rozprawy (Transactions) of the Cracow Acad. [1] xiii. p. 27 ; also 

 Comjptes Rendus, c. p. 350 (1885). 



