234 PHENOMENA DEPENDENT ON MOLECULAR PATHS 90 



ness, in the magnitudes that are independent of the density 

 S, viz. : 



The magnitudes r) l and ?7 2 contained in s are the limiting 

 values which the coefficient of viscosity attains, the one 

 when the vapour or gas is entirely decomposed by dissocia- 

 tion into simple molecules, the other when all the molecules 

 are combined in pairs. Of these limiting values we can at 

 most assume rj 1 as known, since we put for it the value of 

 the coefficient of viscosity which has been observed for 

 carbonic acid under such circumstances that it may pass 

 without question as a perfect gas, i.e. under low pressure at 

 a sufficiently high temperature. The other limiting value, 

 >? 2 , is, however, not known, and must, indeed, itself be calcu- 

 lated from the observations. Therefore e, which depends on 

 the ratio of these two limiting values, is unknown, and the 

 same is true, therefore, of both the magnitudes A and B 

 which occur in the formula. 



These two magnitudes A and B depend not only on s, 

 which according to the theory ought to be constant, but 

 also on the density D of the completely dissociated gas. 

 What unit we are to choose for the numerical value of this 

 magnitude is on the theory entirely at our pleasure ; we can 

 take the density of any gas whatever, such as atmospheric 

 air, as unit, or we are equally at liberty to refer the number 

 to the density of water as unit ; it is only necessary that for 

 both 8 and D one and the same unit should be chosen. 

 Since Warburg and von Babo refer the density 8 of 

 the carbonic acid which they have themselves measured to 

 that of water as unit, we must also choose this unit for D 

 if we wish to employ directly the values which they have 

 tabulated and which they have denoted by s. 



Doubts, however, may be raised whether D is not 

 variable with the pressure, and as to the law of variation, if 

 variation does occur. Since D denotes the density of the gas 

 when all its molecules are decomposed into simple ones, we 

 might be inclined to assume that the gas then behaves as a 



