Professor Sir James Dewar 



[Jan. 22 



Plate 5. — Effect of Peese>'ce of Hydrogex o:^ the Volatility 

 AT 20"^ Abs. of Nitrogen and Carbonic Oxide. 



measure of the pressure was taken to determine the amount of the 

 evaporated material obtained, with the surprising result, in the case of 

 the hydrogen mixtures, that there was nothing practically to measure. 

 Under the high exhaust of the pump, therefore, not only had the 

 uncondensable hydrogen been removed, but also the condensable 

 constituent ; nitrogen, or carbonic oxide, had suffered a forced vola- 

 tility in presence of the hydrogen, although naturally witliout 

 admixture of hydrogen, the nitrogen or carbonic oxide was regained 

 unaltered. In view of this, it seems difficult to resist the conclusion 

 that some sort of molecular aggregation had taken place at the very 

 low temperature, between hydrogen and the carbonic oxide and 

 nitrogen respectively. 



The composition of the gases in this experiment is of course very 

 different from that dealt with in determining the small proportion of 

 air uncondensable. There is, therefore, only a limited application to 

 that case. Thus the amount of condensed nitrogen employed is only 

 one-fourth of the total amount of hydrogen, whereas in the experiment 

 with ordinary air there is five hundred thousand times more nitrogen 

 than hydrogen ; the amount of " forced volatiUty " will therefore 

 probably be so small as to escape measurement, as the difference to 

 be detected would have to be measured on a total of one atmosphere 

 in the gauge. Further, it is possible to make a correct estimate of 

 the amount of occlusion occurring on condensation to 20° Abs. in 

 the case shown in the last table, because the actual composition of 



