8 
Suppose first we compare A and B. Let t and t 1 be the 
lengths of the columns. Then 
W 0 R^ = W 0 R*F (1). 
Now compare B with C, whence, if 0 and 0 1 be the lengths 
of the columns, we shall have 
WoR«^ = WoRW (2) 
i 
From equations (1) we have R = 7 
1 x ' k p — q 
0l Q 
and from equation (2) we have R = r 
but these two values of R ought to be the same, so we 
ought to have the equation. 
t l —t _ 0 1 — Q 
p—q q — r 
The different tints I used consisted of BaS0 4 and lamp 
black. Tint A consisted of lamp black 0'012 grms. BaS0 4 
10 grms. Tint B contained twice the above quantity of 
lamp black to the same quantity of BaS0 4 , and tint C 
contained four times the quantity of lamp black to the 
same quantity of BaS0 4 . The absorbing medium I used 
consisted of water containing a minute quantity of lamp 
black in suspension. This, as I have before shown, behaves 
nearly the same with regard to the absorption of light as if 
the carbon were in solution. 
A comparison of tint A with tint B gave R=&‘ 3 . Tint 
A compared with tint C gave R=&' 346 . Tint B compared 
with C gave R— ' 4 . Inasmuch as k is a fraction, these 
three values of R will not differ much. The experimental 
enquiry is difficult, and the following defect is likely to 
have some influence on the result. The grey powders were 
mixed with a few drops of water and pieces of cardboard 
covered with the paint so obtained, and then dried. But if 
we take an intimate mixture of two powders, and make it 
into a paint with oil or water, the gravity of the two powders 
being different, and the fluid medium imparting a certain 
degree of mobility to the particles, there will be a tendency 
