Mr Bevan, On the Joule- Thomson Effect. LS3 



dioxide, however, the constants thus used do not agree with the 

 constants as deduced from van der Waals' equation *. 



Equation (7) does not therefore lead to very encouraging 

 results as far as agreement with experimental results is con- 

 cerned. We can see that the same inconsistency applies to 

 equation (4) from which equation (7) is derived. 



Equation (4) is equivalent to 



pv = R0 + p0<j> 1 (0), 



or p(v-0<f> 1 (0)) = R0 (8), 



and for a given value of the curves representing pv and p are 

 straight lines. But we know from Amagat's results that the pv 

 and p curves are straight lines approximately only in the case of 

 hydrogen. In all other cases these curves are convex to the 

 pressure axis. 



Our equation is therefore inconsistent with the results of 

 experiment, and we must examine closely the assumptions we 

 have made to find where the inconsistency has been introduced. 



First of all let us consider the assumption that G p is, as far as 



the small term representing the Joule-Thomson effect is concerned, 



independent of the variation of pressure. 



dC 

 For gases which are with difficulty liquefiable the value of ~~ 



is undoubtedly very small at pressures which are not very great, 

 and at temperatures which are not very low. We may assume in 

 virtue of the relation 



dp)e 30V 



rif 1 \ 



that the ratio of -k^\ to G p is of the order of the departure of 



the gas from Boyle's Law. In the small term therefore represent- 

 ing the cooling effect, the result of assuming C p independent of 

 the pressure will be to introduce errors of the second order in the 

 quantities involved. This cannot account for the differences 

 between Amagat's pv and p curves and the curves represented 

 by equation (8), which are of the first order in the same small 

 quantities. 



The other assumption on which we have proceeded is that the 

 Joule-Thomson effect is proportional to the pressure fall from one 

 side of the plug to the other. 



The experimental results on which this assumption is founded 

 are not in good agreement, and it appears that this assumption is. 

 not justifiable. 



* See Callendar, Phil. Mag., Jan. 1903, p. 59. 



9—3 



