904 
generally. There is. indeed, a remarkable difference for invariable 
molecules, i.e. for such for which the quantity 46 does not change. 
= ; 8 Uk Uk : 4 
Phen the quantity we and ——-—— = 3. In this case @ is nak 
t ‘ Vlim 
9 
4, bat = If there exists a rectilinear diameter for substances with 
such molecules. y = - 
bo | 
Perfect equality or almost perfect equality can therefore, only be 
expected for substances for which 6 greatly varies with the volume. 
Thus for substances for which s is about equal to 3.77, the value 
of y is about equal to 0.8 or 0.9. The following remarks are the 
result of my investigation to get more certainty about this question. 
According to the formula: 
Ve 
OC ln ays 
Ulim 
is 
by 
r—— = YS 
Drum 
or 
by s 3? 64 (f—1) 
ZEE 
Olin : : 7 F937 87 
or 
by f-1 8 8 
en En me En Me El Y 3 
Dim 5 9 sr 
8 8 
Now the thought has foreed itself upon me to put a aa 4 
and hence also 
by j—l 
bias = 3 
8 8 Oene, Ge 
The relation Rea igs 1 Ore == mn is satisfied for substances for 
oy Lomas I © 
9 
which 4 is constant. Then 7s = 8, and as we saw above ¢ =e 
For substances with variable value of 8, 7s << 8 and p decreases, 
1 
but comparatively slowly. Not before rs = 7 a would be = 1, 
x 
and for substances for which 7s has this value, the rule 
2(L 4+ yee 
