906 
(sr) << 8, but for by far the majority of the closely investigated 
ne 2(i+y) 
substances (sr) > 7 = For them” must be > 1. But the pos- 
Perey a de. ae ME 
sibility of ———~ <1 is not excluded even for normal substances. 
Yet we should not lose sight of the fact that it has not yet been 
investigated in how far the existence of quasi-association has influence 
on the rule of the rectilinear diameter. A close investigation about 
the value of (rs) for different substances, and comparison of the 
value of y following from this with the experimental data is, there- 
fore, very desirable. 
b, 
If from the knowledge of the value of —’ we want to determine 
lim 
the value of.7, the given relations are not sufficient for an accurate 
by Dy ar ao 
Bhim ee: 
bg ITA 
bim an r 27 
B 
indeed, holds. As, however, the value of the factor of — is not accurately 
5 
known, and as we only know that this factor is smaller than 1, and the 
determination. The relation 
or 
b . 
smaller as —*- is greater, we can only give a value for 7 below 
) lim 
b 
: 5 : we : g : : 
which it must remain. Thus for —~=2 the value of r is below 
lim 
3 | 8 8 8 
— = 2,12. Already with the formula — < 1, or r<-, or # <= 
Ve ; rs s s 
- 8 bg 
we arrive at the given value for r. If we, namely, put s = = ae 
lim 
Ee 3 
r EE =e 
Blim 
Only in the case that 5 is invariable, the sign < must be replaced 
by the sign =. But even for such a great variability of 5 that the 
if : 
value of rs would have decreased to 7 oe r would still amount to 
