296 PHENOMENA DEPENDENT ON MOLECULAR PATHS 108 



are so greatly variable with the temperature as is the vis- 

 cosity of mercury vapour, according to Koch's formula. It 

 cannot, therefore, be surprising if for mercury vapour the 

 ratio of the observed value of f to the calculated value of the 

 product Tjc, given by Will In er, is 3'15, and therefore nearly 

 twice as great as the ratio 1*6027 required by the theory. 

 This great excess of the calculated value of the ratio may 

 be taken as a confirmation of the view put out in 92, that 

 Koch's numbers are too small because a portion of the 

 mercury vapour may have been condensed into the liquid 

 form in the capillary tube. The behaviour of mercury may 

 also perhaps be explained by the division of the molecules 

 into single atoms ( 54) not being complete at 203. But 

 we are in no way compelled to see any obstacle to the theory 

 in this single exception, so long as it is not proved on 

 surer grounds to be an exception. 



The excellent agreement of the calculated and observed 

 values shown by all other gases justifies in us, on the contrary, 

 the conviction that the accuracy of the theoretically deduced 

 relation between the conductivity and viscosity of gases is 

 no longer to be doubted, and that we may take it as proved 

 that a gas has the same conductivity for every kind of 

 energy. 



From this result of theory we see finally that viscosity, 

 diffusion, and conduction of gases depend in the same way 

 on the free path of the gaseous particles, and that each of 

 these three phenomena may be employed to determine the 

 value of the molecular free path. 



