188 PHENOMENA DEPENDENT ON MOLECULAR PATHS 77 



increase as the pressure rises, yet all these experimental 

 results admit of other explanations as well. We may 

 therefore conclude, as the united testimony of all the 

 observations, that within wide limits of the pressure the 

 internal friction of gases is variable with the pressure either 

 not at all or only in a very slight degree. 



78. Numerical Values of the Free Path and 

 Collision-frequency of Particles of Air 



This confirmation of a predicted fact constitutes a 

 brilliant success for the kinetic theory. We may assure 

 ourselves by it that we are not moving in the doubtful 

 region of hypothesis, but on the sure ground of experiment, 

 when we employ the results that have been thus far ob- 

 tained to widen our knowledge of the molecular motion and 

 to investigate the direct properties of molecules. 



The formula established in 75 for the coefficient of 

 viscosity of a gas, 



TJ = P LG, 



now that it is justified by experiment, assumes a no smaller 

 importance than the analogous formula 



p = iTrplP 



that holds for the pressure ( 11 and 27) ; for just as we 

 employ the latter to calculate the absolute value of the 

 molecular speed, we ca,n use the former to determine the 

 value of the molecular free path. The formula enables us 

 to calculate in absolute measure the numerical value of this 

 length, which is apparently inaccessible to observation, 

 when the internal friction of a gas measured in absolute 

 units is known and the value of the molecular speed is 

 obtained from the pressure and density. 



Before we follow Maxwell in this important step we 

 must point out that the formula for the viscosity is capable 

 of a slight improvement. It was deduced in 75 by a con- 



the U.S. Geological Survey, No. 54, ' Measurement of High Temperatures,' 

 Washington 1889. De Keen, Bull.de I'Acad. de Belgique, 1888 [3] xvi. 

 p. 195. 



