286 THE REALITIES OF MODERN SCIENCE 



is removed by this cooling. At the upper end of the 

 chamber C there are some molecules of the vapor, but 

 these are much reduced in velocity by the cooling and 

 but few of them find their way back, against the stream, 

 into the region A. The stream of molecules from c 

 spreads out somewhat as a result of the haphazard 

 motions of its individual molecules. The mass motion 

 given to the gas molecules may therefore be lateral in 

 part, in which case they may bounce a few times from 

 the walls of C, but the motion is nevertheless in the 

 general direction of the exit, B. The lateral spreading 

 of the vapor molecules means collisions with the walls 

 and hence a transfer of part of their energy to the 

 cooler molecules of the walls. In this way but few 

 of them reach the exit C. 



In commercial form this pump has been developed 

 along lines proposed by Langmuir. The limit of the 

 vacua to be obtained by its use does not appear to 

 be inherent in the pump itself but to depend upon the 

 character of the walls of the vessels, particularly that 

 which is to be exhausted. We recognize that these 

 walls are not smooth and solid but are a granular 

 structure of glass molecules. Within the spaces be- 

 tween these molecules there may be numbers of gas 

 molecules. These diffuse into what we usually con- 

 sider the free space of the vessel, but do so compara- 

 tively slowly because of the network of glass molecules 

 past which they must thread their way. Gas, which 

 is thus occluded by the walls of the vessel, is probably 

 the real limit of the vacuum which may be obtained, 

 for otherwise, sooner or later, every gas molecule of 

 the vessel at A would find its way into the chamber C. 



