77 VISCOSITY OF GASES 185 



perimental proof of the independence of the friction on the 

 pressure by use of a method like Coulomb's. Maxwell's 

 apparatus differs from mine by an improvement that is 

 apparently only slight, but in reality is extremely important. 1 

 Between the three oscillating discs fixed to a common axis, 

 as well as above and below them, Maxwell placed four 

 fixed discs at distances from them that were everywhere the 

 same and as little as possible. The first effect of this is to 

 cause the friction of the air to exert on the oscillating discs 

 a greater resistance, and one therefore more easily measured. 

 A second advantage of Maxwell's method consists in the 

 mathematical theory leading to a final formula which gives 

 the viscosity -coefficient directly, while mine led first of all 

 to the square root of the coefficient, to a magnitude, that 

 is, the error of which would be doubled on squaring. For 

 these reasons Maxwell's numbers show the constancy of 

 the coefficient of viscosity much better than mine. The 

 following table contains the mean values given in the much 

 fuller table 2 drawn up by him. The values of the pressures 

 are given in inches of mercury, and opposite each value of 

 the observed logarithmic decrement is put the corresponding 

 value which is calculated on the assumption of the con- 

 stancy of the coefficient of viscosity : 



The agreement of these numbers proves the correctness of 

 the law required by the kinetic theory, viz. that the co- 

 efficient of viscosity of air is independent of its pressure 

 within the limits of pressure from 1 to -fa atmosphere. 



Maxwell further experimented on the friction of 

 hydrogen and carbonic acid. For a great number of 

 other gases a similarity of behaviour was proved by the 



1 See my memoir on this method (Pogg. Ann. 1871, cxliii. p. 14). 



2 Phil. Trans. 1866, clvi. p. 266, Table IV. 



