170 PHENOMENA DEPENDENT ON MOLECULAR PATHS 72 



cules like it, perhaps many thousands, without obstruction, 

 before it finally collides with another, yet the absolute value 

 of the free path, as measured in our usual units, may be an 

 insignificantly small magnitude on account of the extreme 

 minuteness of the distance of molecular separation. 



It is on these circumstances that the possibility of ex- 

 plaining the apparent contradiction between the phenomena 

 mentioned at the beginning of Part II. ( 61) and the theory 

 developed in Part I. depends. The fundamental hypo- 

 thesis of the latter of molecules rapidly moving in straight 

 paths does not exclude the other, viz. that the lengths of 

 these paths, measured in ordinary units, are very small ; it 

 is sufficient that they should be large in comparison with 

 the dimensions and the distances apart of the molecules 

 for the justness of Clausius' assumption to be upheld; 

 for then it is allowable to suppose that the length of path 

 which a molecule passes over in a straight line between 

 successive collisions may be considered as infinitely great 

 in comparison with the curved path it traverses while casu- 

 ally within the sphere of action of another. 



Whether this mode of getting over the difficulty is right 

 and based on actuality is for observation to decide. We 

 turn, therefore, to the experimental methods of actually 

 measuring molecular free paths, and to this end we proceed 

 to the consideration of the viscosity, diffusion, and heat- 

 conduction in gases. 



