﻿Volume, and Temperature of Rarefied Gases. 319 



An interesting calculation may be made with this value. 

 If the gas under examination were air and not hydrogen, 

 then a great quantity of air would have condensed on the 

 glass when the mercury was raised in the gauge to take the 

 reading. If the curve (No. I.) were applied to this value, we 

 should obtain the reading of the vacuum '0000022 millim. 

 This curve, however, was obtained for air at greater pressure 

 than the above, and the condensation would be still greater 

 at "000076 millim., so that this is probably the highest 

 vacuum that has yet been measured. 



These results prove that hydrogen does not in any way 

 condense on to a glass surface on rise of pressure, and that 

 therefore in working a McLeod gauge with hydrogen all the 

 gas in the gauge is measured and none lost by surface-con- 

 densation. The McLeod gauge is accordingly trustworthy to a 

 certain extent when working with absolutely pure hydrogen. 



With carbon dioxide, however, the case is different. Several 

 series of factors of exhaustion were obtained with this gas, 

 and it was found that they steadily increase in value as the 

 exhaustion proceeds. This is owing to condensed carbon 

 dioxide coming off the walls of the apparatus. It seems even 

 possible to reach to a pressure which is equal to the " vapour- 

 pressure " of the carbon dioxide which is condensed on the 

 walls. For in a particular experiment the factors of exhaustion 

 increased steadily till they reached *997, and 188 pumps more 

 failed to make the slightest difference in the vacuum, as 

 evidenced by the electric discharge in the vacuum-tube and 

 by the gauge-readings. It would appear therefore that the 

 pressure in the apparatus just equalled the pressure of the 

 condensed carbon dioxide, and that as fast as the pressure 

 was lowered by more pumping it was restored to its equilibrium 

 by more carbon dioxide escaping from the pores of the con- 

 taining walls. The vacuum was not very high, the apparent 

 pressure being about one millionth of an atmosphere. 



We found the converse of this when admitting measured 

 volumes of carbon dioxide into a vacuum. In these experi- 

 ments a gauge-reading was first taken and the pv calculated ; 

 then a known volume of carbon dioxide also calculated to pv 

 was admitted. This last value multiplied by the ratio which 

 the gauge volume bears to the rest of the apparatus will give 

 the amount of carbon dioxide taken up by the gauge. This 

 added to the previous value of pv read in the gauge ought to 

 agree with the gauge-reading taken after the admission of 

 the carbon dioxide. But we found that the measured pv was 

 always smaller than that calculated. This shows therefore 

 that some of the carbon dioxide condenses. 



