570 JAMES CLERK MAXWELL. 



TABLE OF MOLECULAR DATA, 



Divided into three ranks, according to the degree of accuracy 

 with which the Quantities are known. 



Hydrogen. Oxygen, 



Mass of Molecule (Hydrogen = 1) 1 16 14 22 



flank 



I. Velocity (of mean square) metres) Ig59 m ^ ^ 



per second at C . . ) 



Mean path, tenth-metres . . 965 560 482 379 



Bank 



II. Collisions in a second (millions) 17,750 7646 9489 9720 

 Rank 



III. Diameter, tenth-metres . . 5 '8 7 '6 8 '3 9 '3 



Mass, twenty-fifth-grammes . 46 736 644 1012 



30th June 1877. 



DEAR GARNETT ... I have been considering diffusion 

 of gases, and the method of separating heavy gases from light 

 ones, and I find it hopeless to do it by gravity, but if a tube 

 10 cm. long with two bulbs, and the straight part stuffed with 

 cotton-wool, were filled with equal volumes of H and C0 2 , and 

 spun 100 times round per second for about half an hour, then 

 the ratio of C0 2 to H by volume would be greater in A than in 

 B by about T ^-Q-, which is measurable. I have got a new light 

 about equilibrium of temperature in two different gases. Let 

 forces having potentials act on the molecules of two gases, but 

 differently on each. Let the potential of forces acting on the 

 gas a be zero in the region A and very large in B, diminishing 

 continuously in the stratum C. Let the potential for gas b be 

 zero in B and very great in A, diminishing continuously in C. 

 Then the region A will contain the gas A nearly pure, and B 

 the gas B nearly pure, and in the stratum C there will be en- 

 counters between the two kinds of molecules. By Boltzmann 

 and Watson the average kinetic energy of a single molecule is 

 the same throughout the whole vessel. Hence the condition of 

 thermal equilibrium between two gases (not mixed, but kept 

 pure though in contact) is that the mean kinetic energy is the 

 same in each. And it is difficult to see where this method 

 breaks down when applied to solids. 



I find the electric conductivity of air supposed of conducting 

 spheres to be 



