J. Dewar — Thermal Evolution of Gases. 299 



A manometer attached to the vessel containing the charcoal 

 showed, on shutting off the air current, that during the early 

 part of the saturation the absorption was so effective as to give 

 practically no measurable mercury pressure. As soon as the 

 absorption was ended, and a current began to pass slowly over 

 the charcoal, the composition of the air leaving the charcoal 

 showed 98 per cent nitrogen. After the current of air had 

 passed for half an hour, the total gas occluded in the charcoal 

 was expelled by taking the vessel in which it had been treated 

 out of the liquid air, and allowing the temperature to rise to 

 15° C. 



The gas, which was rapidly expelled, measured 5*7 liters, and 

 contained 56 per cent of oxygen. If the saturated charcoal 

 before heating up was subjected for an hour to the action of 

 an air-pump, capable of giving a steady exhaustion of 5 mm , no 

 difference was effected in the oxygen percentage of the evolved 

 gas. The same experiment was repeated with this variation, 

 that, instead of the air current having the pressure of the 

 atmosphere, it was kept below one-tenth of an atmosphere. 

 In this experiment, 4' 8 liters were expelled on heating up, and 

 the percentage of oxygen was 58. Then a further repetition 

 was made with an air current supplied at a pressure not exceed- 

 ing 5 mm of mercury. After 3 hours' treatment, the charcoal, 

 on heating to 15° C, gave 4-J- liters of 57 per cent oxygen. 

 From these experiments it follows that the tension of the 

 occluded gases, at the temperature of liquid air, must be very 

 small, and thus the use of low temperatures, combined with 

 charcoal, introduces a new and greatly improved means of 

 getting high vacua, which in the future may be found suscep- 

 tible of important practical applications. These experiments 

 are quite conclusive as to the practical constancy of the mean 

 composition of the air gases occluded in the charcoal (subject 

 to the conditions aforesaid), and they further show that wide 

 changes in the pressure of the air current has little or no effect 

 in altering the proportions. In another experiment, the vessel 

 containing the saturated charcoal, instead of being allowed to 

 rise rapidly in temperature, was transferred to a vacuum vessel, 

 in which a little liquid air was placed, in order that the tem- 

 perature might rise slowly, and thereby enable the successive 

 liters of gas given off to be collected separately and analyzed. 



This experiment gave the following results : — 



Oxygen per cent. 



First liter 18*5 



Second " 30-6 



Third " 530 



Fourth " _ 72-0 



Fifth " 79-0 



Sixth " 84-0 



