184 PHENOMENA DEPENDENT ON MOLECULAR PATHS 77 



a given volume of gas, and P the excess of pressure in inches 

 which causes the flow ; D is an auxiliary quantity, propor- 

 tional to the viscosity-coefficient 77, which is used in the 

 calculation, and is constant like that coefficient ; the added 

 values of ?/, referred to centimetres and seconds as units, 

 have been calculated from the mean values of D. $ is the 

 length of the transpiration tube in inches. 



Other series of experiments show an equally excellent 

 agreement, but unfortunately the calculation of the viscosity- 

 coefficients cannot be made in absolute measure from the 

 results given. 1 



Simultaneously with this second memoir of mine there 

 appeared, as has already been indicated, a paper by 

 Maxwell 2 himself, in which he too brought forward ex- 



1 The remarkable slowness with which, according to the numbers published, 

 air and other gases flow through narrow tubes has called a remark from 

 Stefan (Wiener Sitzungsber. 1874, Ixix. p. 713), which, though of no further 

 importance for our theory, deserves mention. The adhesion of flat plates 

 together is only an apparent attraction ; the cause of the phenomenon is the 

 slowness with which the air penetrates into the narrow interval between the 

 plates ; they hang therefore together for a long time by reason of the pressure 

 exerted by the air on their external surfaces. 



2 Proc. Boy. Soc. 1866, xv. p. 14 ; Phil. Trans. 1866, clvi. p. 249. 



