36 Proceedings of the Royal Society of Edinburgh. [Sess. 
... o- - cr n - T — ^ = RT I 2 f P -( da ) ^ 
0 0T Jo A\0T )p V 
RT 2 
A 
If we denote by A the divergence between observed and calculated 
values of the heat of adsorption, and by 2 the expression cr 0 — T 
have 
d(r 0 
0T ’ 
we 
A = - A U-T 
^0cr 
1 0A 
0T7 A da 
approx. 
-{ AET ^'^»- AS )xs; 
- W a a AS )I c ‘ pr *“- 
1 0A 
Since A2 is unknown, we cannot evaluate 7 . 
A da 
proceed by another step. From (20) 
RT 2 
( 20 ) 
0A\ 
da )t 
= J> T 2 t a l°gp # j_ 9A a 02A id A 
0T A da da 2 1 da 
. . . ( 20 ') 
We may, however, 
02AA 
da 2 
=RT2 aiog£ _ isa + a 
0T 
A da 
d_n 0AV 
1 0A da \A da ) 
A da ^ 1 0A 
A da 
(21) 
1 3A A 3 1 3A / 1 3A 
A da da A da / A da 
0A 
If the second term in (21) is negligible compared with — , we may 
OCv 
evaluate 
1 0A 
A . Assuming the term negligible for the moment, we may 
A OQj 
apply (21) to the first three observations in Table I, the only place where 
the tendency to increase or decrease of the differences A is clearly indicated. 
Here it is presumably due to the upward bend in the heat curve shown 
in the figure. 
Neglecting the doubtful value of A, we have from the second and third 
values, when a = 82, 
i 0A 
•022/37 = *355— . 
A da 
I ^=-0017-. 
A da 
From (20'), 
•051 =(28 - A2) -0017. 
A!2 = - 3 approx. 
