228 PROCEEDINGS OF THE AMERICAN ACADEMY. 



whence it follows that 



1 mc 



'^ ~ 4 77^2 y 



or for any two different gases, 



miC, mnC 



But, since for any two gases at constant temperature, 



the above proportion may be simplified to the following equation : 



Hl = Vi .M. (1) 



from which it is evident that the relative mean cross-sections of the 

 molecules of the two gases are readily calculated from their molecular 

 weights and viscosity coefficients. It was thought by us that monatomic 

 molecules might prove to be much smaller than polyatomic ones, since it 

 seems a 'priori not improbable that the spaces between the atoms of the 

 latter are large in comparison with the dimensions of the atoms them- 

 selves. The experiments to be here described show, however, that no 

 marked distinction exists between monatomic and polyatomic gases in 

 this respect. 



Experiments on the viscosity of mercury vapor, and especially on the 

 effect of temperature upon it, have been already made by S. Koch, * who 

 calculated that at 300° the volume of mercury molecule is 4.4 as great 

 as that of the hydrogen molecule. As this calculation was not based on 

 direct comparative experiments made by passing the two gases through 

 the same capillary, but was an indirect one involving the measurements 

 of different experimenters, and the dimensions of the capillaries used by 

 them, it seemed desirable to subject the matter to further investigation in 

 the direct manner indicated. Moreover, the author does not discuss the 

 significance of his result in its bearing on the relative magnitude of atoms 

 and molecules. 



The method used by us in determining the relative viscosity consisted 



* Wied. Ann. Pliys. Clicm., XIX. 857, 1883. 



