﻿Volume, and Temperature of Rarefied Gases. 323 



Having thus satisfactorily proved that this method of 

 determining the thermal expansion of hydrogen is capable of 

 giving accurate results, we proceeded to apply it to de- 

 terming the expansions of oxygen and nitrogen at reduced 

 pressures. 



In the case of oxygen the gas was made by heating per- 

 manganate of potassium, it was washed in caustic potash and 

 dried by means of sulphuric acid and phosphorus pentoxide. 

 Oxygen proved to be an exceptionally difficult gas to deal 

 with, as it was found impossible to determine its purity by the 

 electric discharge in the vacuum-tube, as the presence of dust 

 or other combustible material led to the formation of carbon 

 dioxide. This method therefore was of no use, and the purity 

 of the oxygen was determined by the result obtained, as will 

 be seen further on. 



The coefficient of expansion of oxygen was found to be 

 from very careful experiments : — 



5*1 millim. 

 5-3 „ 



26L? 



260' 



J_ 



262" 



These results are very fairly good, and being at such high 

 pressures as 4 and 5 millim. may be said to demonstrate the 

 absence of carbon dioxide, especially inasmuch as they give 

 evidence of oxygen being too expansible while carbon dioxide 

 is the opposite. Each of these results, too, was obtained after 

 a fresh filling of the apparatus, which in each case was most 

 carefully done. Therefore the proof that they are measure- 

 ments of the expansion of pure oxygen may be regarded as 

 conclusive. As the results next obtained gave evidence that 

 with decrease of pressure the coefficient of expansion still 

 further increased, and as on the slightest sign of presence of 

 carbon dioxide the apparatus was most carefully filled again 

 with oxygen, there is no doubt that the expansion of pure 

 oxygen was in each case measured. The presence of a small 

 quantity of carbon dioxide would materially alter the results, 

 as the value of the coefficient of expansion of this gas is so 

 small at the pressures here dealt with. 



The anomaly in the expansion of oxygen, first noticed by 

 Bohr, was brought very strikingly under our notice in working 

 with oxygen. The vacuum was at one time 1*4 millim., when, 

 on a stroke of the pump being taken, the remaining gas, now 

 at a pressure of about *75 millim., behaved in a most extra- 

 ordinary fashion. For example, the two gauges E and F, 



